From t3kcit at gmail.com  Mon May 16 20:31:47 2016
From: t3kcit at gmail.com (Andreas Mueller)
Date: Mon, 16 May 2016 20:31:47 -0400
Subject: [scikit-learn] Welcome to the moved scikit-learn mailing list.
Message-ID: <573A6673.3050702@gmail.com>

Hi everybody!

Welcome to the new scikit-learn mailing list!

Cheers,
Andy

From mathieu at mblondel.org  Tue May 17 00:35:11 2016
From: mathieu at mblondel.org (Mathieu Blondel)
Date: Tue, 17 May 2016 13:35:11 +0900
Subject: [scikit-learn] Welcome to the moved scikit-learn mailing list.
In-Reply-To: <573A6673.3050702@gmail.com>
References: <573A6673.3050702@gmail.com>
Message-ID: <CAOKSrLzzVdDRQJmAOakVeSciLZ=-J92cnwK0y67w+4Zm0v3iUA@mail.gmail.com>

Great news Andy!

How did you manage to transfer all subscribers?

Mathieu

On Tue, May 17, 2016 at 9:31 AM, Andreas Mueller <t3kcit at gmail.com> wrote:

> Hi everybody!
>
> Welcome to the new scikit-learn mailing list!
>
> Cheers,
> Andy
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160517/abb8a469/attachment.html>

From robert.kern at gmail.com  Tue May 17 03:23:53 2016
From: robert.kern at gmail.com (Robert Kern)
Date: Tue, 17 May 2016 08:23:53 +0100
Subject: [scikit-learn] Welcome to the moved scikit-learn mailing list.
In-Reply-To: <573A6673.3050702@gmail.com>
References: <573A6673.3050702@gmail.com>
Message-ID: <CAF6FJis5YQomWiXSGQSivxeaaFSsOxMxpTSNmYX86tRDQGForw@mail.gmail.com>

I've emailed the GMane administrator to let them know about the new list.
Thanks!

On Tue, May 17, 2016 at 1:31 AM, Andreas Mueller <t3kcit at gmail.com> wrote:

> Hi everybody!
>
> Welcome to the new scikit-learn mailing list!
>
> Cheers,
> Andy
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>



-- 
Robert Kern
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160517/b1a6131a/attachment.html>

From gael.varoquaux at normalesup.org  Tue May 17 03:46:08 2016
From: gael.varoquaux at normalesup.org (Gael Varoquaux)
Date: Tue, 17 May 2016 09:46:08 +0200
Subject: [scikit-learn] Welcome to the moved scikit-learn mailing list.
In-Reply-To: <CAF6FJis5YQomWiXSGQSivxeaaFSsOxMxpTSNmYX86tRDQGForw@mail.gmail.com>
References: <573A6673.3050702@gmail.com>
 <CAF6FJis5YQomWiXSGQSivxeaaFSsOxMxpTSNmYX86tRDQGForw@mail.gmail.com>
Message-ID: <20160517074608.GB765255@phare.normalesup.org>

Good point. Thanks!

Ga?l

On Tue, May 17, 2016 at 08:23:53AM +0100, Robert Kern wrote:
> I've emailed the GMane administrator to let them know about the new list.
> Thanks!

> On Tue, May 17, 2016 at 1:31 AM, Andreas Mueller <t3kcit at gmail.com> wrote:

>     Hi everybody!

>     Welcome to the new scikit-learn mailing list!

>     Cheers,
>     Andy
>     _______________________________________________
>     scikit-learn mailing list
>     scikit-learn at python.org
>     https://mail.python.org/mailman/listinfo/scikit-learn
-- 
    Gael Varoquaux
    Researcher, INRIA Parietal
    NeuroSpin/CEA Saclay , Bat 145, 91191 Gif-sur-Yvette France
    Phone:  ++ 33-1-69-08-79-68
    http://gael-varoquaux.info            http://twitter.com/GaelVaroquaux

From olivier.grisel at ensta.org  Tue May 17 06:54:25 2016
From: olivier.grisel at ensta.org (Olivier Grisel)
Date: Tue, 17 May 2016 12:54:25 +0200
Subject: [scikit-learn] Welcome to the moved scikit-learn mailing list.
In-Reply-To: <573A6673.3050702@gmail.com>
References: <573A6673.3050702@gmail.com>
Message-ID: <CAFvE7K4Oe1nF2rc3V2YN7nBU4GGAKZn_t56xw0YA5BxBGx-+-Q@mail.gmail.com>

Thanks for making it happen!

-- 
Olivier

From t3kcit at gmail.com  Tue May 17 08:31:00 2016
From: t3kcit at gmail.com (Andreas Mueller)
Date: Tue, 17 May 2016 08:31:00 -0400
Subject: [scikit-learn] Welcome to the moved scikit-learn mailing list.
In-Reply-To: <CAFvE7K4Oe1nF2rc3V2YN7nBU4GGAKZn_t56xw0YA5BxBGx-+-Q@mail.gmail.com>
References: <573A6673.3050702@gmail.com>
 <CAFvE7K4Oe1nF2rc3V2YN7nBU4GGAKZn_t56xw0YA5BxBGx-+-Q@mail.gmail.com>
Message-ID: <CADxzQorhM9TCwYhUhdeOiJDohV_GrHrR9KxpuZO_Gs1wmiChTg@mail.gmail.com>

Hm we need to update the websites. Maybe the stable one, too.
I kind of forgot about that.

Mathieu: via the mailman web interface.
Though I have no idea how I extracted the users from it lol. I had a
notebook recording all subscribers for counting them and their institutes.
I have no recollection of how I got that. Theoretically you can get that
via the mail interface but apparently that's not how I did it.

Sent from phone. Please excuse spelling and brevity.
On May 17, 2016 06:56, "Olivier Grisel" <olivier.grisel at ensta.org> wrote:

> Thanks for making it happen!
>
> --
> Olivier
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160517/5bbf52c2/attachment-0001.html>

From mike at selik.org  Tue May 17 10:05:52 2016
From: mike at selik.org (Michael Selik)
Date: Tue, 17 May 2016 14:05:52 +0000
Subject: [scikit-learn] Welcome to the moved scikit-learn mailing list.
In-Reply-To: <CADxzQorhM9TCwYhUhdeOiJDohV_GrHrR9KxpuZO_Gs1wmiChTg@mail.gmail.com>
References: <573A6673.3050702@gmail.com>
 <CAFvE7K4Oe1nF2rc3V2YN7nBU4GGAKZn_t56xw0YA5BxBGx-+-Q@mail.gmail.com>
 <CADxzQorhM9TCwYhUhdeOiJDohV_GrHrR9KxpuZO_Gs1wmiChTg@mail.gmail.com>
Message-ID: <CADqi794Exxqj56twN0s3sZfcV7FRYiKZX6hiamP3gDYSWW_zxg@mail.gmail.com>

I've just now discovered that I'm on this list (pleasant surprise). Why
didn't I receive emails previously, was there little activity?

On Tue, May 17, 2016, 8:32 AM Andreas Mueller <t3kcit at gmail.com> wrote:

> Hm we need to update the websites. Maybe the stable one, too.
> I kind of forgot about that.
>
> Mathieu: via the mailman web interface.
> Though I have no idea how I extracted the users from it lol. I had a
> notebook recording all subscribers for counting them and their institutes.
> I have no recollection of how I got that. Theoretically you can get that
> via the mail interface but apparently that's not how I did it.
>
> Sent from phone. Please excuse spelling and brevity.
> On May 17, 2016 06:56, "Olivier Grisel" <olivier.grisel at ensta.org> wrote:
>
>> Thanks for making it happen!
>>
>> --
>> Olivier
>> _______________________________________________
>> scikit-learn mailing list
>> scikit-learn at python.org
>> https://mail.python.org/mailman/listinfo/scikit-learn
>>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160517/8797fea1/attachment.html>

From t3kcit at gmail.com  Tue May 17 10:13:43 2016
From: t3kcit at gmail.com (Andreas Mueller)
Date: Tue, 17 May 2016 10:13:43 -0400
Subject: [scikit-learn] Welcome to the moved scikit-learn mailing list.
In-Reply-To: <CADqi794Exxqj56twN0s3sZfcV7FRYiKZX6hiamP3gDYSWW_zxg@mail.gmail.com>
References: <573A6673.3050702@gmail.com>
 <CAFvE7K4Oe1nF2rc3V2YN7nBU4GGAKZn_t56xw0YA5BxBGx-+-Q@mail.gmail.com>
 <CADxzQorhM9TCwYhUhdeOiJDohV_GrHrR9KxpuZO_Gs1wmiChTg@mail.gmail.com>
 <CADqi794Exxqj56twN0s3sZfcV7FRYiKZX6hiamP3gDYSWW_zxg@mail.gmail.com>
Message-ID: <573B2717.2060100@gmail.com>




On 05/17/2016 10:05 AM, Michael Selik wrote:
>
> I've just now discovered that I'm on this list (pleasant surprise). 
> Why didn't I receive emails previously, was there little activity?
>
>
Maybe you had a filter sending it to spam or an obscure folder? ;)

From jendunning93 at gmail.com  Tue May 17 10:26:06 2016
From: jendunning93 at gmail.com (Jennifer Dunning)
Date: Tue, 17 May 2016 09:26:06 -0500
Subject: [scikit-learn] Welcome to the moved scikit-learn mailing list.
In-Reply-To: <573B2717.2060100@gmail.com>
References: <573A6673.3050702@gmail.com>
 <CAFvE7K4Oe1nF2rc3V2YN7nBU4GGAKZn_t56xw0YA5BxBGx-+-Q@mail.gmail.com>
 <CADxzQorhM9TCwYhUhdeOiJDohV_GrHrR9KxpuZO_Gs1wmiChTg@mail.gmail.com>
 <CADqi794Exxqj56twN0s3sZfcV7FRYiKZX6hiamP3gDYSWW_zxg@mail.gmail.com>
 <573B2717.2060100@gmail.com>
Message-ID: <CACVMLH_Zo-=4u2dFc5fYAKfZhBA7H4mv2TneYTwU2sMWjzRiVQ@mail.gmail.com>

I previously unsubscribed, but now it seems I've been added again. I
decided I would prefer to read the messages online. There was just too much
activity for my inbox.

How can I unsubscribe again?
On May 17, 2016 9:14 AM, "Andreas Mueller" <t3kcit at gmail.com> wrote:

>
>
>
> On 05/17/2016 10:05 AM, Michael Selik wrote:
>
>>
>> I've just now discovered that I'm on this list (pleasant surprise). Why
>> didn't I receive emails previously, was there little activity?
>>
>>
>> Maybe you had a filter sending it to spam or an obscure folder? ;)
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160517/368a5968/attachment.html>

From mhancock743 at gmail.com  Tue May 17 10:28:33 2016
From: mhancock743 at gmail.com (Matt)
Date: Tue, 17 May 2016 10:28:33 -0400
Subject: [scikit-learn] Welcome to the moved scikit-learn mailing list.
In-Reply-To: <573B2717.2060100@gmail.com>
References: <573A6673.3050702@gmail.com>
 <CAFvE7K4Oe1nF2rc3V2YN7nBU4GGAKZn_t56xw0YA5BxBGx-+-Q@mail.gmail.com>
 <CADxzQorhM9TCwYhUhdeOiJDohV_GrHrR9KxpuZO_Gs1wmiChTg@mail.gmail.com>
 <CADqi794Exxqj56twN0s3sZfcV7FRYiKZX6hiamP3gDYSWW_zxg@mail.gmail.com>
 <573B2717.2060100@gmail.com>
Message-ID: <573B2A91.7070205@gmail.com>

Maybe not? I have never received emails from this list until the switch, 
either. I don't appear to have any old emails in my spam from sklearn 
previous to the switch.

I'm not complaining, because I'm an avid sklearn user, and I don't doubt 
that I probably willingly submitted my e-mail to something sklearn 
related in the past. I am curious, however, as to why I'm getting the 
e-mails only now, and where my email was pulled from.

-Matt


On 05/17/2016 10:13 AM, Andreas Mueller wrote:

> Maybe you had a filter sending it to spam or an obscure folder? ;)
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn


From robert.kern at gmail.com  Tue May 17 10:34:48 2016
From: robert.kern at gmail.com (Robert Kern)
Date: Tue, 17 May 2016 15:34:48 +0100
Subject: [scikit-learn] Welcome to the moved scikit-learn mailing list.
In-Reply-To: <CACVMLH_Zo-=4u2dFc5fYAKfZhBA7H4mv2TneYTwU2sMWjzRiVQ@mail.gmail.com>
References: <573A6673.3050702@gmail.com>
 <CAFvE7K4Oe1nF2rc3V2YN7nBU4GGAKZn_t56xw0YA5BxBGx-+-Q@mail.gmail.com>
 <CADxzQorhM9TCwYhUhdeOiJDohV_GrHrR9KxpuZO_Gs1wmiChTg@mail.gmail.com>
 <CADqi794Exxqj56twN0s3sZfcV7FRYiKZX6hiamP3gDYSWW_zxg@mail.gmail.com>
 <573B2717.2060100@gmail.com>
 <CACVMLH_Zo-=4u2dFc5fYAKfZhBA7H4mv2TneYTwU2sMWjzRiVQ@mail.gmail.com>
Message-ID: <CAF6FJivUHa+-2K8yF=FMigisfu8JsWUjJmEHP5=BZBn9Zw6ihg@mail.gmail.com>

On Tue, May 17, 2016 at 3:26 PM, Jennifer Dunning <jendunning93 at gmail.com>
wrote:
>
> I previously unsubscribed, but now it seems I've been added again. I
decided I would prefer to read the messages online. There was just too much
activity for my inbox.
>
> How can I unsubscribe again?

Follow the link for this mailing list at the bottom of these messages. The
unsubscribe form is at the bottom of the page.

https://mail.python.org/mailman/listinfo/scikit-learn

--
Robert Kern
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160517/13e94b93/attachment-0001.html>

From sbenthall at gmail.com  Tue May 17 12:21:45 2016
From: sbenthall at gmail.com (Sebastian Benthall)
Date: Tue, 17 May 2016 09:21:45 -0700
Subject: [scikit-learn] Welcome to the moved scikit-learn mailing list.
In-Reply-To: <CADxzQorhM9TCwYhUhdeOiJDohV_GrHrR9KxpuZO_Gs1wmiChTg@mail.gmail.com>
References: <573A6673.3050702@gmail.com>
 <CAFvE7K4Oe1nF2rc3V2YN7nBU4GGAKZn_t56xw0YA5BxBGx-+-Q@mail.gmail.com>
 <CADxzQorhM9TCwYhUhdeOiJDohV_GrHrR9KxpuZO_Gs1wmiChTg@mail.gmail.com>
Message-ID: <CAEYE9Of-n9aq6a6r7Bh99gbjMz2uiGgPryef2DXO4+agqR98=Q@mail.gmail.com>

Would you be willing to share the notebook? It sounds interesting.
On May 17, 2016 5:33 AM, "Andreas Mueller" <t3kcit at gmail.com> wrote:

> Hm we need to update the websites. Maybe the stable one, too.
> I kind of forgot about that.
>
> Mathieu: via the mailman web interface.
> Though I have no idea how I extracted the users from it lol. I had a
> notebook recording all subscribers for counting them and their institutes.
> I have no recollection of how I got that. Theoretically you can get that
> via the mail interface but apparently that's not how I did it.
>
> Sent from phone. Please excuse spelling and brevity.
> On May 17, 2016 06:56, "Olivier Grisel" <olivier.grisel at ensta.org> wrote:
>
>> Thanks for making it happen!
>>
>> --
>> Olivier
>> _______________________________________________
>> scikit-learn mailing list
>> scikit-learn at python.org
>> https://mail.python.org/mailman/listinfo/scikit-learn
>>
>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160517/ba601706/attachment.html>

From t3kcit at gmail.com  Tue May 17 13:25:15 2016
From: t3kcit at gmail.com (Andreas Mueller)
Date: Tue, 17 May 2016 13:25:15 -0400
Subject: [scikit-learn] Welcome to the moved scikit-learn mailing list.
In-Reply-To: <CAEYE9Of-n9aq6a6r7Bh99gbjMz2uiGgPryef2DXO4+agqR98=Q@mail.gmail.com>
References: <573A6673.3050702@gmail.com>
 <CAFvE7K4Oe1nF2rc3V2YN7nBU4GGAKZn_t56xw0YA5BxBGx-+-Q@mail.gmail.com>
 <CADxzQorhM9TCwYhUhdeOiJDohV_GrHrR9KxpuZO_Gs1wmiChTg@mail.gmail.com>
 <CAEYE9Of-n9aq6a6r7Bh99gbjMz2uiGgPryef2DXO4+agqR98=Q@mail.gmail.com>
Message-ID: <573B53FB.1020805@gmail.com>

I filter by ending in "edu", take the part after the "@", drop 
subdomains and do a value_count ;)


On 05/17/2016 12:21 PM, Sebastian Benthall wrote:
>
> Would you be willing to share the notebook? It sounds interesting.
>
> On May 17, 2016 5:33 AM, "Andreas Mueller" <t3kcit at gmail.com 
> <mailto:t3kcit at gmail.com>> wrote:
>
>     Hm we need to update the websites. Maybe the stable one, too.
>     I kind of forgot about that.
>
>     Mathieu: via the mailman web interface.
>     Though I have no idea how I extracted the users from it lol. I had
>     a notebook recording all subscribers for counting them and their
>     institutes. I have no recollection of how I got that.
>     Theoretically you can get that via the mail interface but
>     apparently that's not how I did it.
>
>     Sent from phone. Please excuse spelling and brevity.
>
>     On May 17, 2016 06:56, "Olivier Grisel" <olivier.grisel at ensta.org
>     <mailto:olivier.grisel at ensta.org>> wrote:
>
>         Thanks for making it happen!
>
>         --
>         Olivier
>         _______________________________________________
>         scikit-learn mailing list
>         scikit-learn at python.org <mailto:scikit-learn at python.org>
>         https://mail.python.org/mailman/listinfo/scikit-learn
>
>
>     _______________________________________________
>     scikit-learn mailing list
>     scikit-learn at python.org <mailto:scikit-learn at python.org>
>     https://mail.python.org/mailman/listinfo/scikit-learn
>
>
>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160517/7193f4db/attachment.html>

From sam at samnicholls.net  Tue May 17 13:46:11 2016
From: sam at samnicholls.net (Sam Nicholls)
Date: Tue, 17 May 2016 18:46:11 +0100
Subject: [scikit-learn] Welcome to the moved scikit-learn mailing list.
In-Reply-To: <573B53FB.1020805@gmail.com>
References: <573A6673.3050702@gmail.com>
 <CAFvE7K4Oe1nF2rc3V2YN7nBU4GGAKZn_t56xw0YA5BxBGx-+-Q@mail.gmail.com>
 <CADxzQorhM9TCwYhUhdeOiJDohV_GrHrR9KxpuZO_Gs1wmiChTg@mail.gmail.com>
 <CAEYE9Of-n9aq6a6r7Bh99gbjMz2uiGgPryef2DXO4+agqR98=Q@mail.gmail.com>
 <573B53FB.1020805@gmail.com>
Message-ID: <CA++Nc6hs2Q9ZQTe7Y8kTry-skP4yBZsLnhLeyO1KGTdfhEXy5g@mail.gmail.com>

What about the *.ac.uk folks across the pond?!

On Tue, May 17, 2016 at 6:25 PM, Andreas Mueller <t3kcit at gmail.com> wrote:

> I filter by ending in "edu", take the part after the "@", drop subdomains
> and do a value_count ;)
>
>
>
> On 05/17/2016 12:21 PM, Sebastian Benthall wrote:
>
> Would you be willing to share the notebook? It sounds interesting.
> On May 17, 2016 5:33 AM, "Andreas Mueller" <t3kcit at gmail.com> wrote:
>
>> Hm we need to update the websites. Maybe the stable one, too.
>> I kind of forgot about that.
>>
>> Mathieu: via the mailman web interface.
>> Though I have no idea how I extracted the users from it lol. I had a
>> notebook recording all subscribers for counting them and their institutes.
>> I have no recollection of how I got that. Theoretically you can get that
>> via the mail interface but apparently that's not how I did it.
>>
>> Sent from phone. Please excuse spelling and brevity.
>> On May 17, 2016 06:56, "Olivier Grisel" <olivier.grisel at ensta.org> wrote:
>>
>>> Thanks for making it happen!
>>>
>>> --
>>> Olivier
>>> _______________________________________________
>>> scikit-learn mailing list
>>> scikit-learn at python.org
>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>
>>
>> _______________________________________________
>> scikit-learn mailing list
>> scikit-learn at python.org
>> https://mail.python.org/mailman/listinfo/scikit-learn
>>
>>
>
> _______________________________________________
> scikit-learn mailing listscikit-learn at python.orghttps://mail.python.org/mailman/listinfo/scikit-learn
>
>
>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160517/8e675ab1/attachment.html>

From t3kcit at gmail.com  Tue May 17 13:59:42 2016
From: t3kcit at gmail.com (Andreas Mueller)
Date: Tue, 17 May 2016 13:59:42 -0400
Subject: [scikit-learn] Welcome to the moved scikit-learn mailing list.
In-Reply-To: <CA++Nc6hs2Q9ZQTe7Y8kTry-skP4yBZsLnhLeyO1KGTdfhEXy5g@mail.gmail.com>
References: <573A6673.3050702@gmail.com>
 <CAFvE7K4Oe1nF2rc3V2YN7nBU4GGAKZn_t56xw0YA5BxBGx-+-Q@mail.gmail.com>
 <CADxzQorhM9TCwYhUhdeOiJDohV_GrHrR9KxpuZO_Gs1wmiChTg@mail.gmail.com>
 <CAEYE9Of-n9aq6a6r7Bh99gbjMz2uiGgPryef2DXO4+agqR98=Q@mail.gmail.com>
 <573B53FB.1020805@gmail.com>
 <CA++Nc6hs2Q9ZQTe7Y8kTry-skP4yBZsLnhLeyO1KGTdfhEXy5g@mail.gmail.com>
Message-ID: <573B5C0E.20903@gmail.com>

The NSF doesn't care about you, I was told ;)


On 05/17/2016 01:46 PM, Sam Nicholls wrote:
> What about the *.ac.uk <http://ac.uk> folks across the pond?!
>
> On Tue, May 17, 2016 at 6:25 PM, Andreas Mueller <t3kcit at gmail.com 
> <mailto:t3kcit at gmail.com>> wrote:
>
>     I filter by ending in "edu", take the part after the "@", drop
>     subdomains and do a value_count ;)
>
>
>
>     On 05/17/2016 12:21 PM, Sebastian Benthall wrote:
>>
>>     Would you be willing to share the notebook? It sounds interesting.
>>
>>     On May 17, 2016 5:33 AM, "Andreas Mueller" <t3kcit at gmail.com
>>     <mailto:t3kcit at gmail.com>> wrote:
>>
>>         Hm we need to update the websites. Maybe the stable one, too.
>>         I kind of forgot about that.
>>
>>         Mathieu: via the mailman web interface.
>>         Though I have no idea how I extracted the users from it lol.
>>         I had a notebook recording all subscribers for counting them
>>         and their institutes. I have no recollection of how I got
>>         that. Theoretically you can get that via the mail interface
>>         but apparently that's not how I did it.
>>
>>         Sent from phone. Please excuse spelling and brevity.
>>
>>         On May 17, 2016 06:56, "Olivier Grisel"
>>         <olivier.grisel at ensta.org <mailto:olivier.grisel at ensta.org>>
>>         wrote:
>>
>>             Thanks for making it happen!
>>
>>             --
>>             Olivier
>>             _______________________________________________
>>             scikit-learn mailing list
>>             scikit-learn at python.org <mailto:scikit-learn at python.org>
>>             https://mail.python.org/mailman/listinfo/scikit-learn
>>
>>
>>         _______________________________________________
>>         scikit-learn mailing list
>>         scikit-learn at python.org <mailto:scikit-learn at python.org>
>>         https://mail.python.org/mailman/listinfo/scikit-learn
>>
>>
>>
>>     _______________________________________________
>>     scikit-learn mailing list
>>     scikit-learn at python.org <mailto:scikit-learn at python.org>
>>     https://mail.python.org/mailman/listinfo/scikit-learn
>
>
>     _______________________________________________
>     scikit-learn mailing list
>     scikit-learn at python.org <mailto:scikit-learn at python.org>
>     https://mail.python.org/mailman/listinfo/scikit-learn
>
>
>
>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160517/e16e0991/attachment-0001.html>

From pczzy13 at gmail.com  Tue May 17 07:25:31 2016
From: pczzy13 at gmail.com (zhang peng)
Date: Tue, 17 May 2016 19:25:31 +0800
Subject: [scikit-learn] Welcome to the moved scikit-learn mailing list.
In-Reply-To: <CAFvE7K4Oe1nF2rc3V2YN7nBU4GGAKZn_t56xw0YA5BxBGx-+-Q@mail.gmail.com>
References: <573A6673.3050702@gmail.com>
 <CAFvE7K4Oe1nF2rc3V2YN7nBU4GGAKZn_t56xw0YA5BxBGx-+-Q@mail.gmail.com>
Message-ID: <CE0FF01F-66CF-433A-8306-850A9E2E02EC@gmail.com>

New mailing list works fine

???? iPhone

> ? 2016?5?17????6:54?Olivier Grisel <olivier.grisel at ensta.org> ???
> 
> Thanks for making it happen!
> 
> -- 
> Olivier
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn

From sson800 at aucklanduni.ac.nz  Tue May 17 16:46:59 2016
From: sson800 at aucklanduni.ac.nz (Shen S)
Date: Wed, 18 May 2016 08:46:59 +1200
Subject: [scikit-learn] Welcome to the moved scikit-learn mailing list.
In-Reply-To: <573A6673.3050702@gmail.com>
References: <573A6673.3050702@gmail.com>
Message-ID: <CAAswceiC1s-2D4E4uGj-M4SxEbxC0mRGb3iZJ9Tm-r0Pgns9jg@mail.gmail.com>

Hi Andy,

How can I unsubscribe from the list? Would you please give me a hand?

Thank you very much.

Cheers,
    SShen

On 17 May 2016 at 12:31, Andreas Mueller <t3kcit at gmail.com> wrote:

> Hi everybody!
>
> Welcome to the new scikit-learn mailing list!
>
> Cheers,
> Andy
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160518/7d8e5455/attachment.html>

From jianjin78 at gmail.com  Tue May 17 17:02:16 2016
From: jianjin78 at gmail.com (Jianjian Jin)
Date: Tue, 17 May 2016 17:02:16 -0400
Subject: [scikit-learn] unsubscribe from the mailing list
Message-ID: <CAB_Qp9tjh-OdgWBpWOk1=aVtB+506V_V+xY1zo51bCX0H=AotA@mail.gmail.com>

Hi,

   Looks like there are too many emails, so I'll have to unsubscribe to the
mailing list, would you help me with that? Thanks a lot.

best

Jianjian
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160517/b5d93e1c/attachment.html>

From matthieu.brucher at gmail.com  Tue May 17 17:44:37 2016
From: matthieu.brucher at gmail.com (Matthieu Brucher)
Date: Tue, 17 May 2016 22:44:37 +0100
Subject: [scikit-learn] unsubscribe from the mailing list
In-Reply-To: <CAB_Qp9tjh-OdgWBpWOk1=aVtB+506V_V+xY1zo51bCX0H=AotA@mail.gmail.com>
References: <CAB_Qp9tjh-OdgWBpWOk1=aVtB+506V_V+xY1zo51bCX0H=AotA@mail.gmail.com>
Message-ID: <CAHCaCk+g=_jdvLc66ROO0O_w2PNukfX2SkDPGKAgq=nCg=rtSw@mail.gmail.com>

Jsut read the end of the email.

2016-05-17 22:02 GMT+01:00 Jianjian Jin <jianjin78 at gmail.com>:

> Hi,
>
>    Looks like there are too many emails, so I'll have to unsubscribe to
> the mailing list, would you help me with that? Thanks a lot.
>
> best
>
> Jianjian
>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>


-- 
Information System Engineer, Ph.D.
Blog: http://matt.eifelle.com
LinkedIn: http://www.linkedin.com/in/matthieubrucher
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160517/cd5ef815/attachment.html>

From mathieu at mblondel.org  Wed May 18 03:15:03 2016
From: mathieu at mblondel.org (Mathieu Blondel)
Date: Wed, 18 May 2016 16:15:03 +0900
Subject: [scikit-learn] Welcome to the moved scikit-learn mailing list.
In-Reply-To: <CADxzQorhM9TCwYhUhdeOiJDohV_GrHrR9KxpuZO_Gs1wmiChTg@mail.gmail.com>
References: <573A6673.3050702@gmail.com>
 <CAFvE7K4Oe1nF2rc3V2YN7nBU4GGAKZn_t56xw0YA5BxBGx-+-Q@mail.gmail.com>
 <CADxzQorhM9TCwYhUhdeOiJDohV_GrHrR9KxpuZO_Gs1wmiChTg@mail.gmail.com>
Message-ID: <CAOKSrLyHkEkoaHTW8sFP19zqy+3qji5fUYSEuadUELLxtO8X6A@mail.gmail.com>

On Tue, May 17, 2016 at 9:31 PM, Andreas Mueller <t3kcit at gmail.com> wrote:

> Mathieu: via the mailman web interface.
> Though I have no idea how I extracted the users from it lol. I had a
> notebook recording all subscribers for counting them and their institutes.
> I have no recollection of how I got that.
>
So you used an old copy of the subscriber list? This would mean that recent
subscribers were not transferred to the new mailing-list...

Mathieu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160518/1c50cbbc/attachment.html>

From t3kcit at gmail.com  Wed May 18 09:07:29 2016
From: t3kcit at gmail.com (Andy)
Date: Wed, 18 May 2016 09:07:29 -0400
Subject: [scikit-learn] Welcome to the moved scikit-learn mailing list.
In-Reply-To: <CAOKSrLyHkEkoaHTW8sFP19zqy+3qji5fUYSEuadUELLxtO8X6A@mail.gmail.com>
References: <573A6673.3050702@gmail.com>
 <CAFvE7K4Oe1nF2rc3V2YN7nBU4GGAKZn_t56xw0YA5BxBGx-+-Q@mail.gmail.com>
 <CADxzQorhM9TCwYhUhdeOiJDohV_GrHrR9KxpuZO_Gs1wmiChTg@mail.gmail.com>
 <CAOKSrLyHkEkoaHTW8sFP19zqy+3qji5fUYSEuadUELLxtO8X6A@mail.gmail.com>
Message-ID: <573C6911.5050106@gmail.com>



On 05/18/2016 03:15 AM, Mathieu Blondel wrote:
>
> On Tue, May 17, 2016 at 9:31 PM, Andreas Mueller <t3kcit at gmail.com 
> <mailto:t3kcit at gmail.com>> wrote:
>
>     Mathieu: via the mailman web interface.
>     Though I have no idea how I extracted the users from it lol. I had
>     a notebook recording all subscribers for counting them and their
>     institutes. I have no recollection of how I got that.
>
> So you used an old copy of the subscriber list? This would mean that 
> recent subscribers were not transferred to the new mailing-list...
Yes, where old means about a week old.
But I sent a message to the old list that it moved and people that 
didn't get a welcome to the new list should subscribe.
I hoped that was sufficient.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160518/66b4e637/attachment-0001.html>

From felsig at gmail.com  Thu May 19 02:32:45 2016
From: felsig at gmail.com (=?UTF-8?Q?Roxana_Ordo=C3=B1ez?=)
Date: Thu, 19 May 2016 06:32:45 +0000 (UTC)
Subject: [scikit-learn] Please unsubscribe me.
In-Reply-To: <573A6673.3050702@gmail.com>
References: <573A6673.3050702@gmail.com>
Message-ID: <B28B75F9AE57BE71.EDCCA46A-4EFA-48E6-AC9D-7ED382FBDAC5@mail.outlook.com>

Please unsubscribe me. Thanks.?

Rocky (Roxana)
305.807.9506
sent.from.a.tiny.computer.




On Mon, May 16, 2016 at 5:32 PM -0700, "Andreas Mueller" <t3kcit at gmail.com> wrote:










Hi everybody!

Welcome to the new scikit-learn mailing list!

Cheers,
Andy
_______________________________________________
scikit-learn mailing list
scikit-learn at python.org
https://mail.python.org/mailman/listinfo/scikit-learn





-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160519/e62d6f05/attachment.html>

From benjaminaaronblumer at gmail.com  Thu May 19 02:35:43 2016
From: benjaminaaronblumer at gmail.com (Benjamin Blumer)
Date: Wed, 18 May 2016 23:35:43 -0700
Subject: [scikit-learn] Please unsubscribe me.
In-Reply-To: <B28B75F9AE57BE71.EDCCA46A-4EFA-48E6-AC9D-7ED382FBDAC5@mail.outlook.com>
References: <573A6673.3050702@gmail.com>
 <B28B75F9AE57BE71.EDCCA46A-4EFA-48E6-AC9D-7ED382FBDAC5@mail.outlook.com>
Message-ID: <CAN2qn0BdPjoNqGDTeUtM1tdbK_p2X=V8g=e9NOkxM2WGJ7nvYQ@mail.gmail.com>

Could we change the signature so that the link to the mailman page says
"unsubscribe & other options"?
On 18 May 2016 11:33 p.m., "Roxana Ordo?ez" <felsig at gmail.com> wrote:

> Please unsubscribe me. Thanks.
>
> Rocky (Roxana)
> 305.807.9506
> sent.from.a.tiny.computer.
>
>
>
>
> On Mon, May 16, 2016 at 5:32 PM -0700, "Andreas Mueller" <t3kcit at gmail.com
> > wrote:
>
> Hi everybody!
>>
>> Welcome to the new scikit-learn mailing list!
>>
>> Cheers,
>> Andy
>> _______________________________________________
>> scikit-learn mailing listscikit-learn at python.orghttps://mail.python.org/mailman/listinfo/scikit-learn
>>
>>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160518/e7bbe9a7/attachment.html>

From jmwoloso at asu.edu  Tue May 24 12:43:59 2016
From: jmwoloso at asu.edu (Jason Wolosonovich)
Date: Tue, 24 May 2016 09:43:59 -0700
Subject: [scikit-learn] Digest Working?
Message-ID: <574484CF.1070107@asu.edu>

Hey All,


I haven't gotten any emails from the mailing list for quite a few days. 
I suppose this could be possible if no one has any questions, but 
usually there is /something/. I guess I'll find out if this message 
shows up, lol.



-Jason
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160524/6f2e7685/attachment.html>

From nfliu at uw.edu  Tue May 24 13:10:17 2016
From: nfliu at uw.edu (Nelson Liu)
Date: Tue, 24 May 2016 17:10:17 +0000
Subject: [scikit-learn] Digest Working?
In-Reply-To: <574484CF.1070107@asu.edu>
References: <574484CF.1070107@asu.edu>
Message-ID: <CALoLHMLT7JjaEkVAfDKA3CWt0bA7=fzwAJzwDhwpu6UOhALVMg@mail.gmail.com>

Hi Jason,
In the future, you could also check the list archives (
https://mail.python.org/pipermail/scikit-learn/) to see if you missed
anything or if the digest was not updating you appropriately.
Nelson Liu

On Tue, May 24, 2016 at 10:04 AM Jason Wolosonovich <jmwoloso at asu.edu>
wrote:

> Hey All,
>
>
> I haven't gotten any emails from the mailing list for quite a few days. I
> suppose this could be possible if no one has any questions, but usually
> there is *something*. I guess I'll find out if this message shows up, lol.
>
>
>
> -Jason
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160524/385c3106/attachment.html>

From ragvrv at gmail.com  Wed May 25 20:20:57 2016
From: ragvrv at gmail.com (Raghav R V)
Date: Thu, 26 May 2016 02:20:57 +0200
Subject: [scikit-learn] Change of github handle
Message-ID: <CACmxyDGzCpFsSHdvVRy2VFgHiYs3o=kZiTWOQHWj3=Mj=72Z4Q@mail.gmail.com>

Hi scikit devs,

I'm changing my github handle (hopefully permanently) from "rvraghav93" to
"raghavrv" (and twitter handle to "_raghavrv").

Apologies for the inconvenience!

Thanks!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160526/6c6aa0a1/attachment.html>

From blrstartuphire at gmail.com  Thu May 26 02:08:19 2016
From: blrstartuphire at gmail.com (Startup Hire)
Date: Thu, 26 May 2016 11:38:19 +0530
Subject: [scikit-learn] Fitting Lognormal Distribution
Message-ID: <CAPEF3M-JxUD1OUqeuUbiXR6bGTGomFW=gZtARyJxQ5kMCMmSOA@mail.gmail.com>

Hi all,

Hope you are doing good.

I am working on a project where I need to do the following things:

1. I need to fit a lognormal distribution to a set of values [I know its
lognormal by a simple XY scatter plot in excel]

2. I need to find the intersection of the lognormal distribution so that I
can decide cut-off values based on that.


Can you guide me on (1) and (2) can be achieved in python?

Regards,
Sanant
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160526/5a228df3/attachment.html>

From michael.eickenberg at gmail.com  Thu May 26 02:22:54 2016
From: michael.eickenberg at gmail.com (Michael Eickenberg)
Date: Thu, 26 May 2016 08:22:54 +0200
Subject: [scikit-learn] Fitting Lognormal Distribution
In-Reply-To: <CAPEF3M-JxUD1OUqeuUbiXR6bGTGomFW=gZtARyJxQ5kMCMmSOA@mail.gmail.com>
References: <CAPEF3M-JxUD1OUqeuUbiXR6bGTGomFW=gZtARyJxQ5kMCMmSOA@mail.gmail.com>
Message-ID: <CADxJN64_C9wXmYtsn4uc4o+1W8pa0jzGVd4=69ML5N3mpGh4jg@mail.gmail.com>

Hi Sanant,

On Thursday, May 26, 2016, Startup Hire <blrstartuphire at gmail.com> wrote:

> Hi all,
>
> Hope you are doing good.
>

I would like to think so, but you never know where ML will lead us ...


>
> I am working on a project where I need to do the following things:
>
> 1. I need to fit a lognormal distribution to a set of values [I know its
> lognormal by a simple XY scatter plot in excel]
>

if your distribution is lognormal, why don't you try fitting a gaussian to
the log of the values? is this too unstable?


>
> 2. I need to find the intersection of the lognormal distribution so that I
> can decide cut-off values based on that.
>

what exactly do you mean by intersection?


>
>
> Can you guide me on (1) and (2) can be achieved in python?
>
> Regards,
> Sanant
>


Michael
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160526/b3c6b880/attachment.html>

From blrstartuphire at gmail.com  Thu May 26 02:41:38 2016
From: blrstartuphire at gmail.com (Startup Hire)
Date: Thu, 26 May 2016 12:11:38 +0530
Subject: [scikit-learn] Fitting Lognormal Distribution
In-Reply-To: <CADxJN64_C9wXmYtsn4uc4o+1W8pa0jzGVd4=69ML5N3mpGh4jg@mail.gmail.com>
References: <CAPEF3M-JxUD1OUqeuUbiXR6bGTGomFW=gZtARyJxQ5kMCMmSOA@mail.gmail.com>
 <CADxJN64_C9wXmYtsn4uc4o+1W8pa0jzGVd4=69ML5N3mpGh4jg@mail.gmail.com>
Message-ID: <CAPEF3M8rDdOoJRUT4raLFk_44fXpOeT_3BPLMqX=22iPgkUgEQ@mail.gmail.com>

Hi Michael,

:)


(1)  - I think you are right, how do I fit a normal distribution to the log
of values?

(2)  Intersection ---> Meeting point (s)  . as in where the curves cross
each other (it can be in multiple places too!)


Regards,
Sanant

On Thu, May 26, 2016 at 11:52 AM, Michael Eickenberg <
michael.eickenberg at gmail.com> wrote:

> Hi Sanant,
>
> On Thursday, May 26, 2016, Startup Hire <blrstartuphire at gmail.com> wrote:
>
>> Hi all,
>>
>> Hope you are doing good.
>>
>
> I would like to think so, but you never know where ML will lead us ...
>
>
>>
>> I am working on a project where I need to do the following things:
>>
>> 1. I need to fit a lognormal distribution to a set of values [I know its
>> lognormal by a simple XY scatter plot in excel]
>>
>
> if your distribution is lognormal, why don't you try fitting a gaussian to
> the log of the values? is this too unstable?
>
>
>>
>> 2. I need to find the intersection of the lognormal distribution so that
>> I can decide cut-off values based on that.
>>
>
> what exactly do you mean by intersection?
>
>
>>
>>
>> Can you guide me on (1) and (2) can be achieved in python?
>>
>> Regards,
>> Sanant
>>
>
>
> Michael
>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160526/63efd485/attachment.html>

From vaggi.federico at gmail.com  Thu May 26 02:46:28 2016
From: vaggi.federico at gmail.com (federico vaggi)
Date: Thu, 26 May 2016 06:46:28 +0000
Subject: [scikit-learn] Fitting Lognormal Distribution
In-Reply-To: <CAPEF3M8rDdOoJRUT4raLFk_44fXpOeT_3BPLMqX=22iPgkUgEQ@mail.gmail.com>
References: <CAPEF3M-JxUD1OUqeuUbiXR6bGTGomFW=gZtARyJxQ5kMCMmSOA@mail.gmail.com>
 <CADxJN64_C9wXmYtsn4uc4o+1W8pa0jzGVd4=69ML5N3mpGh4jg@mail.gmail.com>
 <CAPEF3M8rDdOoJRUT4raLFk_44fXpOeT_3BPLMqX=22iPgkUgEQ@mail.gmail.com>
Message-ID: <CAGvd0=itnusU=878Vj3yU5tFH5XNHBSaT=+0jNCzSVfj2WqYFg@mail.gmail.com>

1) The normal distribution is parametrized by standard deviation and mean.
Simply take the mean and standard deviation of the log of your values?

2) Which curves?  You only mentioned a single log normal distribution.

On Thu, 26 May 2016 at 08:42 Startup Hire <blrstartuphire at gmail.com> wrote:

> Hi Michael,
>
> :)
>
>
> (1)  - I think you are right, how do I fit a normal distribution to the
> log of values?
>
> (2)  Intersection ---> Meeting point (s)  . as in where the curves cross
> each other (it can be in multiple places too!)
>
>
> Regards,
> Sanant
>
> On Thu, May 26, 2016 at 11:52 AM, Michael Eickenberg <
> michael.eickenberg at gmail.com> wrote:
>
>> Hi Sanant,
>>
>> On Thursday, May 26, 2016, Startup Hire <blrstartuphire at gmail.com> wrote:
>>
>>> Hi all,
>>>
>>> Hope you are doing good.
>>>
>>
>> I would like to think so, but you never know where ML will lead us ...
>>
>>
>>>
>>> I am working on a project where I need to do the following things:
>>>
>>> 1. I need to fit a lognormal distribution to a set of values [I know its
>>> lognormal by a simple XY scatter plot in excel]
>>>
>>
>> if your distribution is lognormal, why don't you try fitting a gaussian
>> to the log of the values? is this too unstable?
>>
>>
>>>
>>> 2. I need to find the intersection of the lognormal distribution so that
>>> I can decide cut-off values based on that.
>>>
>>
>> what exactly do you mean by intersection?
>>
>>
>>>
>>>
>>> Can you guide me on (1) and (2) can be achieved in python?
>>>
>>> Regards,
>>> Sanant
>>>
>>
>>
>> Michael
>>
>> _______________________________________________
>> scikit-learn mailing list
>> scikit-learn at python.org
>> https://mail.python.org/mailman/listinfo/scikit-learn
>>
>>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160526/91581a5e/attachment-0001.html>

From blrstartuphire at gmail.com  Thu May 26 03:23:31 2016
From: blrstartuphire at gmail.com (Startup Hire)
Date: Thu, 26 May 2016 12:53:31 +0530
Subject: [scikit-learn] Fitting Lognormal Distribution
In-Reply-To: <CAGvd0=itnusU=878Vj3yU5tFH5XNHBSaT=+0jNCzSVfj2WqYFg@mail.gmail.com>
References: <CAPEF3M-JxUD1OUqeuUbiXR6bGTGomFW=gZtARyJxQ5kMCMmSOA@mail.gmail.com>
 <CADxJN64_C9wXmYtsn4uc4o+1W8pa0jzGVd4=69ML5N3mpGh4jg@mail.gmail.com>
 <CAPEF3M8rDdOoJRUT4raLFk_44fXpOeT_3BPLMqX=22iPgkUgEQ@mail.gmail.com>
 <CAGvd0=itnusU=878Vj3yU5tFH5XNHBSaT=+0jNCzSVfj2WqYFg@mail.gmail.com>
Message-ID: <CAPEF3M9CK_2W1L=tGrrSgh_hmXPSJG4u+bB3KUB5edBspWwH=Q@mail.gmail.com>

Hi,

(1) - Thanks. will do that

(2) - I am fitting the distribution for 2 different set of values.. I will
find the distribution as mentioned by you in (1).. But, now having 2
curves, how do i find the meetings point(s) ?

Regards,
Sanant

On Thu, May 26, 2016 at 12:16 PM, federico vaggi <vaggi.federico at gmail.com>
wrote:

> 1) The normal distribution is parametrized by standard deviation and
> mean.  Simply take the mean and standard deviation of the log of your
> values?
>
> 2) Which curves?  You only mentioned a single log normal distribution.
>
> On Thu, 26 May 2016 at 08:42 Startup Hire <blrstartuphire at gmail.com>
> wrote:
>
>> Hi Michael,
>>
>> :)
>>
>>
>> (1)  - I think you are right, how do I fit a normal distribution to the
>> log of values?
>>
>> (2)  Intersection ---> Meeting point (s)  . as in where the curves cross
>> each other (it can be in multiple places too!)
>>
>>
>> Regards,
>> Sanant
>>
>> On Thu, May 26, 2016 at 11:52 AM, Michael Eickenberg <
>> michael.eickenberg at gmail.com> wrote:
>>
>>> Hi Sanant,
>>>
>>> On Thursday, May 26, 2016, Startup Hire <blrstartuphire at gmail.com>
>>> wrote:
>>>
>>>> Hi all,
>>>>
>>>> Hope you are doing good.
>>>>
>>>
>>> I would like to think so, but you never know where ML will lead us ...
>>>
>>>
>>>>
>>>> I am working on a project where I need to do the following things:
>>>>
>>>> 1. I need to fit a lognormal distribution to a set of values [I know
>>>> its lognormal by a simple XY scatter plot in excel]
>>>>
>>>
>>> if your distribution is lognormal, why don't you try fitting a gaussian
>>> to the log of the values? is this too unstable?
>>>
>>>
>>>>
>>>> 2. I need to find the intersection of the lognormal distribution so
>>>> that I can decide cut-off values based on that.
>>>>
>>>
>>> what exactly do you mean by intersection?
>>>
>>>
>>>>
>>>>
>>>> Can you guide me on (1) and (2) can be achieved in python?
>>>>
>>>> Regards,
>>>> Sanant
>>>>
>>>
>>>
>>> Michael
>>>
>>> _______________________________________________
>>> scikit-learn mailing list
>>> scikit-learn at python.org
>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>
>>>
>> _______________________________________________
>> scikit-learn mailing list
>> scikit-learn at python.org
>> https://mail.python.org/mailman/listinfo/scikit-learn
>>
>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160526/4c87fc88/attachment.html>

From vaggi.federico at gmail.com  Thu May 26 03:26:53 2016
From: vaggi.federico at gmail.com (federico vaggi)
Date: Thu, 26 May 2016 07:26:53 +0000
Subject: [scikit-learn] Fitting Lognormal Distribution
In-Reply-To: <CAPEF3M9CK_2W1L=tGrrSgh_hmXPSJG4u+bB3KUB5edBspWwH=Q@mail.gmail.com>
References: <CAPEF3M-JxUD1OUqeuUbiXR6bGTGomFW=gZtARyJxQ5kMCMmSOA@mail.gmail.com>
 <CADxJN64_C9wXmYtsn4uc4o+1W8pa0jzGVd4=69ML5N3mpGh4jg@mail.gmail.com>
 <CAPEF3M8rDdOoJRUT4raLFk_44fXpOeT_3BPLMqX=22iPgkUgEQ@mail.gmail.com>
 <CAGvd0=itnusU=878Vj3yU5tFH5XNHBSaT=+0jNCzSVfj2WqYFg@mail.gmail.com>
 <CAPEF3M9CK_2W1L=tGrrSgh_hmXPSJG4u+bB3KUB5edBspWwH=Q@mail.gmail.com>
Message-ID: <CAGvd0=gtXvk-eKF=g1rBkkV0RnVRX+oUnj3Uk5uVQj-XB0WLfg@mail.gmail.com>

If you are talking about finding the values at which the probability
density functions will have the same value, then you can just write the
equations explicitly and solve in terms of theta1, sigma1 and theta2,
sigma2?


On Thu, 26 May 2016 at 09:23 Startup Hire <blrstartuphire at gmail.com> wrote:

> Hi,
>
> (1) - Thanks. will do that
>
> (2) - I am fitting the distribution for 2 different set of values.. I will
> find the distribution as mentioned by you in (1).. But, now having 2
> curves, how do i find the meetings point(s) ?
>
> Regards,
> Sanant
>
> On Thu, May 26, 2016 at 12:16 PM, federico vaggi <vaggi.federico at gmail.com
> > wrote:
>
>> 1) The normal distribution is parametrized by standard deviation and
>> mean.  Simply take the mean and standard deviation of the log of your
>> values?
>>
>> 2) Which curves?  You only mentioned a single log normal distribution.
>>
>> On Thu, 26 May 2016 at 08:42 Startup Hire <blrstartuphire at gmail.com>
>> wrote:
>>
>>> Hi Michael,
>>>
>>> :)
>>>
>>>
>>> (1)  - I think you are right, how do I fit a normal distribution to the
>>> log of values?
>>>
>>> (2)  Intersection ---> Meeting point (s)  . as in where the curves cross
>>> each other (it can be in multiple places too!)
>>>
>>>
>>> Regards,
>>> Sanant
>>>
>>> On Thu, May 26, 2016 at 11:52 AM, Michael Eickenberg <
>>> michael.eickenberg at gmail.com> wrote:
>>>
>>>> Hi Sanant,
>>>>
>>>> On Thursday, May 26, 2016, Startup Hire <blrstartuphire at gmail.com>
>>>> wrote:
>>>>
>>>>> Hi all,
>>>>>
>>>>> Hope you are doing good.
>>>>>
>>>>
>>>> I would like to think so, but you never know where ML will lead us ...
>>>>
>>>>
>>>>>
>>>>> I am working on a project where I need to do the following things:
>>>>>
>>>>> 1. I need to fit a lognormal distribution to a set of values [I know
>>>>> its lognormal by a simple XY scatter plot in excel]
>>>>>
>>>>
>>>> if your distribution is lognormal, why don't you try fitting a gaussian
>>>> to the log of the values? is this too unstable?
>>>>
>>>>
>>>>>
>>>>> 2. I need to find the intersection of the lognormal distribution so
>>>>> that I can decide cut-off values based on that.
>>>>>
>>>>
>>>> what exactly do you mean by intersection?
>>>>
>>>>
>>>>>
>>>>>
>>>>> Can you guide me on (1) and (2) can be achieved in python?
>>>>>
>>>>> Regards,
>>>>> Sanant
>>>>>
>>>>
>>>>
>>>> Michael
>>>>
>>>> _______________________________________________
>>>> scikit-learn mailing list
>>>> scikit-learn at python.org
>>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>>
>>>>
>>> _______________________________________________
>>> scikit-learn mailing list
>>> scikit-learn at python.org
>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>
>>
>> _______________________________________________
>> scikit-learn mailing list
>> scikit-learn at python.org
>> https://mail.python.org/mailman/listinfo/scikit-learn
>>
>>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160526/b47a2418/attachment.html>

From vaggi.federico at gmail.com  Thu May 26 03:27:46 2016
From: vaggi.federico at gmail.com (federico vaggi)
Date: Thu, 26 May 2016 07:27:46 +0000
Subject: [scikit-learn] Fitting Lognormal Distribution
In-Reply-To: <CAGvd0=gtXvk-eKF=g1rBkkV0RnVRX+oUnj3Uk5uVQj-XB0WLfg@mail.gmail.com>
References: <CAPEF3M-JxUD1OUqeuUbiXR6bGTGomFW=gZtARyJxQ5kMCMmSOA@mail.gmail.com>
 <CADxJN64_C9wXmYtsn4uc4o+1W8pa0jzGVd4=69ML5N3mpGh4jg@mail.gmail.com>
 <CAPEF3M8rDdOoJRUT4raLFk_44fXpOeT_3BPLMqX=22iPgkUgEQ@mail.gmail.com>
 <CAGvd0=itnusU=878Vj3yU5tFH5XNHBSaT=+0jNCzSVfj2WqYFg@mail.gmail.com>
 <CAPEF3M9CK_2W1L=tGrrSgh_hmXPSJG4u+bB3KUB5edBspWwH=Q@mail.gmail.com>
 <CAGvd0=gtXvk-eKF=g1rBkkV0RnVRX+oUnj3Uk5uVQj-XB0WLfg@mail.gmail.com>
Message-ID: <CAGvd0=gsKut+0Ak+uSqZyDnGDHxXRUVr3JKBRPL=JjSnxEU6iQ@mail.gmail.com>

Err, sorry - mu1, mu2, sigma1, sigma2, where mu1, sigma1 are the
mean/standard deviation of the first distribution, and mu2, sigma2 are the
mean and standard deviation of the second distribution.

On Thu, 26 May 2016 at 09:26 federico vaggi <vaggi.federico at gmail.com>
wrote:

> If you are talking about finding the values at which the probability
> density functions will have the same value, then you can just write the
> equations explicitly and solve in terms of theta1, sigma1 and theta2,
> sigma2?
>
>
> On Thu, 26 May 2016 at 09:23 Startup Hire <blrstartuphire at gmail.com>
> wrote:
>
>> Hi,
>>
>> (1) - Thanks. will do that
>>
>> (2) - I am fitting the distribution for 2 different set of values.. I
>> will find the distribution as mentioned by you in (1).. But, now having 2
>> curves, how do i find the meetings point(s) ?
>>
>> Regards,
>> Sanant
>>
>> On Thu, May 26, 2016 at 12:16 PM, federico vaggi <
>> vaggi.federico at gmail.com> wrote:
>>
>>> 1) The normal distribution is parametrized by standard deviation and
>>> mean.  Simply take the mean and standard deviation of the log of your
>>> values?
>>>
>>> 2) Which curves?  You only mentioned a single log normal distribution.
>>>
>>> On Thu, 26 May 2016 at 08:42 Startup Hire <blrstartuphire at gmail.com>
>>> wrote:
>>>
>>>> Hi Michael,
>>>>
>>>> :)
>>>>
>>>>
>>>> (1)  - I think you are right, how do I fit a normal distribution to the
>>>> log of values?
>>>>
>>>> (2)  Intersection ---> Meeting point (s)  . as in where the curves
>>>> cross each other (it can be in multiple places too!)
>>>>
>>>>
>>>> Regards,
>>>> Sanant
>>>>
>>>> On Thu, May 26, 2016 at 11:52 AM, Michael Eickenberg <
>>>> michael.eickenberg at gmail.com> wrote:
>>>>
>>>>> Hi Sanant,
>>>>>
>>>>> On Thursday, May 26, 2016, Startup Hire <blrstartuphire at gmail.com>
>>>>> wrote:
>>>>>
>>>>>> Hi all,
>>>>>>
>>>>>> Hope you are doing good.
>>>>>>
>>>>>
>>>>> I would like to think so, but you never know where ML will lead us ...
>>>>>
>>>>>
>>>>>>
>>>>>> I am working on a project where I need to do the following things:
>>>>>>
>>>>>> 1. I need to fit a lognormal distribution to a set of values [I know
>>>>>> its lognormal by a simple XY scatter plot in excel]
>>>>>>
>>>>>
>>>>> if your distribution is lognormal, why don't you try fitting a
>>>>> gaussian to the log of the values? is this too unstable?
>>>>>
>>>>>
>>>>>>
>>>>>> 2. I need to find the intersection of the lognormal distribution so
>>>>>> that I can decide cut-off values based on that.
>>>>>>
>>>>>
>>>>> what exactly do you mean by intersection?
>>>>>
>>>>>
>>>>>>
>>>>>>
>>>>>> Can you guide me on (1) and (2) can be achieved in python?
>>>>>>
>>>>>> Regards,
>>>>>> Sanant
>>>>>>
>>>>>
>>>>>
>>>>> Michael
>>>>>
>>>>> _______________________________________________
>>>>> scikit-learn mailing list
>>>>> scikit-learn at python.org
>>>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>>>
>>>>>
>>>> _______________________________________________
>>>> scikit-learn mailing list
>>>> scikit-learn at python.org
>>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>>
>>>
>>> _______________________________________________
>>> scikit-learn mailing list
>>> scikit-learn at python.org
>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>
>>>
>> _______________________________________________
>> scikit-learn mailing list
>> scikit-learn at python.org
>> https://mail.python.org/mailman/listinfo/scikit-learn
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160526/d593e036/attachment-0001.html>

From mamunbabu2001 at gmail.com  Thu May 26 05:03:49 2016
From: mamunbabu2001 at gmail.com (Mamun Rashid)
Date: Thu, 26 May 2016 10:03:49 +0100
Subject: [scikit-learn] Digest Subscription
Message-ID: <833806CC-4A4D-4489-B732-0CAA0EC82EA9@gmail.com>

Hi All,
Since the mailing list switch I am receiving individual emails instead of digests. I can not seem to change because for some reason my old password is not working. 
Any comments ?

Thanks,
Mamun

From blrstartuphire at gmail.com  Thu May 26 05:58:57 2016
From: blrstartuphire at gmail.com (Startup Hire)
Date: Thu, 26 May 2016 15:28:57 +0530
Subject: [scikit-learn] Fitting Lognormal Distribution
In-Reply-To: <CAGvd0=gsKut+0Ak+uSqZyDnGDHxXRUVr3JKBRPL=JjSnxEU6iQ@mail.gmail.com>
References: <CAPEF3M-JxUD1OUqeuUbiXR6bGTGomFW=gZtARyJxQ5kMCMmSOA@mail.gmail.com>
 <CADxJN64_C9wXmYtsn4uc4o+1W8pa0jzGVd4=69ML5N3mpGh4jg@mail.gmail.com>
 <CAPEF3M8rDdOoJRUT4raLFk_44fXpOeT_3BPLMqX=22iPgkUgEQ@mail.gmail.com>
 <CAGvd0=itnusU=878Vj3yU5tFH5XNHBSaT=+0jNCzSVfj2WqYFg@mail.gmail.com>
 <CAPEF3M9CK_2W1L=tGrrSgh_hmXPSJG4u+bB3KUB5edBspWwH=Q@mail.gmail.com>
 <CAGvd0=gtXvk-eKF=g1rBkkV0RnVRX+oUnj3Uk5uVQj-XB0WLfg@mail.gmail.com>
 <CAGvd0=gsKut+0Ak+uSqZyDnGDHxXRUVr3JKBRPL=JjSnxEU6iQ@mail.gmail.com>
Message-ID: <CAPEF3M8+_nGU6bk+6ENS2kSUxv3=b_AA8qD67qX=0TABJ+ZczQ@mail.gmail.com>

Thanks

On Thu, May 26, 2016 at 12:57 PM, federico vaggi <vaggi.federico at gmail.com>
wrote:

> Err, sorry - mu1, mu2, sigma1, sigma2, where mu1, sigma1 are the
> mean/standard deviation of the first distribution, and mu2, sigma2 are the
> mean and standard deviation of the second distribution.
>
> On Thu, 26 May 2016 at 09:26 federico vaggi <vaggi.federico at gmail.com>
> wrote:
>
>> If you are talking about finding the values at which the probability
>> density functions will have the same value, then you can just write the
>> equations explicitly and solve in terms of theta1, sigma1 and theta2,
>> sigma2?
>>
>>
>> On Thu, 26 May 2016 at 09:23 Startup Hire <blrstartuphire at gmail.com>
>> wrote:
>>
>>> Hi,
>>>
>>> (1) - Thanks. will do that
>>>
>>> (2) - I am fitting the distribution for 2 different set of values.. I
>>> will find the distribution as mentioned by you in (1).. But, now having 2
>>> curves, how do i find the meetings point(s) ?
>>>
>>> Regards,
>>> Sanant
>>>
>>> On Thu, May 26, 2016 at 12:16 PM, federico vaggi <
>>> vaggi.federico at gmail.com> wrote:
>>>
>>>> 1) The normal distribution is parametrized by standard deviation and
>>>> mean.  Simply take the mean and standard deviation of the log of your
>>>> values?
>>>>
>>>> 2) Which curves?  You only mentioned a single log normal distribution.
>>>>
>>>> On Thu, 26 May 2016 at 08:42 Startup Hire <blrstartuphire at gmail.com>
>>>> wrote:
>>>>
>>>>> Hi Michael,
>>>>>
>>>>> :)
>>>>>
>>>>>
>>>>> (1)  - I think you are right, how do I fit a normal distribution to
>>>>> the log of values?
>>>>>
>>>>> (2)  Intersection ---> Meeting point (s)  . as in where the curves
>>>>> cross each other (it can be in multiple places too!)
>>>>>
>>>>>
>>>>> Regards,
>>>>> Sanant
>>>>>
>>>>> On Thu, May 26, 2016 at 11:52 AM, Michael Eickenberg <
>>>>> michael.eickenberg at gmail.com> wrote:
>>>>>
>>>>>> Hi Sanant,
>>>>>>
>>>>>> On Thursday, May 26, 2016, Startup Hire <blrstartuphire at gmail.com>
>>>>>> wrote:
>>>>>>
>>>>>>> Hi all,
>>>>>>>
>>>>>>> Hope you are doing good.
>>>>>>>
>>>>>>
>>>>>> I would like to think so, but you never know where ML will lead us ...
>>>>>>
>>>>>>
>>>>>>>
>>>>>>> I am working on a project where I need to do the following things:
>>>>>>>
>>>>>>> 1. I need to fit a lognormal distribution to a set of values [I know
>>>>>>> its lognormal by a simple XY scatter plot in excel]
>>>>>>>
>>>>>>
>>>>>> if your distribution is lognormal, why don't you try fitting a
>>>>>> gaussian to the log of the values? is this too unstable?
>>>>>>
>>>>>>
>>>>>>>
>>>>>>> 2. I need to find the intersection of the lognormal distribution so
>>>>>>> that I can decide cut-off values based on that.
>>>>>>>
>>>>>>
>>>>>> what exactly do you mean by intersection?
>>>>>>
>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> Can you guide me on (1) and (2) can be achieved in python?
>>>>>>>
>>>>>>> Regards,
>>>>>>> Sanant
>>>>>>>
>>>>>>
>>>>>>
>>>>>> Michael
>>>>>>
>>>>>> _______________________________________________
>>>>>> scikit-learn mailing list
>>>>>> scikit-learn at python.org
>>>>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>>>>
>>>>>>
>>>>> _______________________________________________
>>>>> scikit-learn mailing list
>>>>> scikit-learn at python.org
>>>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>>>
>>>>
>>>> _______________________________________________
>>>> scikit-learn mailing list
>>>> scikit-learn at python.org
>>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>>
>>>>
>>> _______________________________________________
>>> scikit-learn mailing list
>>> scikit-learn at python.org
>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>
>>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160526/10cafc11/attachment.html>

From donkey-hotei at cryptolab.net  Thu May 26 06:46:40 2016
From: donkey-hotei at cryptolab.net (donkey-hotei at cryptolab.net)
Date: Thu, 26 May 2016 12:46:40 +0200
Subject: [scikit-learn] partial_fit implementation for IsolationForest
Message-ID: <e3d4a8743d92e96d3627ce4947271646@cryptolab.net>

hello scikit-learn devs,

After following the work on IsolationForest so far and testing on a 
real-world problem here we've found this model to be very promising for 
anomaly detection. However, at present, IsolationForest only fits data 
in batch even while it may be well suited to incremental on-line 
learning since one could subsample recent history and older estimators 
can be dropped progressively.

I'd like to contribute this feature, but being new to ML and 
scikit-learn I'm curious how I should start making a quick & dirty 
version to see how this may work. Are there other good examples where 
one could see the difference between .fit and .partial_fit in other 
models?

thanks
isaak y.

From arthur.mensch at inria.fr  Thu May 26 07:32:36 2016
From: arthur.mensch at inria.fr (Arthur Mensch)
Date: Thu, 26 May 2016 13:32:36 +0200
Subject: [scikit-learn] partial_fit implementation for IsolationForest
In-Reply-To: <e3d4a8743d92e96d3627ce4947271646@cryptolab.net>
References: <e3d4a8743d92e96d3627ce4947271646@cryptolab.net>
Message-ID: <CAGmDsiMq-s4kwGjVNguESka13dcJGyGs_aZqbNxBBcfFHj3frQ@mail.gmail.com>

Hi Isaac,

You may have a look at MiniBatchKMeans and MiniBatchDictionaryLearning that
both proposes this API. At the moment, you should fit a single mini batch
to the estimator using partial_fit, and update the inner attributes
accordingly. During the first partial_fit, you should take care of various
memory allocation that are needed by the estimator.

Please fill free to create a pull request whenever you think your code is
ready for review.

Good luck!
Le 26 mai 2016 13:14, <donkey-hotei at cryptolab.net> a ?crit :

> hello scikit-learn devs,
>
> After following the work on IsolationForest so far and testing on a
> real-world problem here we've found this model to be very promising for
> anomaly detection. However, at present, IsolationForest only fits data in
> batch even while it may be well suited to incremental on-line learning
> since one could subsample recent history and older estimators can be
> dropped progressively.
>
> I'd like to contribute this feature, but being new to ML and scikit-learn
> I'm curious how I should start making a quick & dirty version to see how
> this may work. Are there other good examples where one could see the
> difference between .fit and .partial_fit in other models?
>
> thanks
> isaak y.
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160526/50bf8e58/attachment-0001.html>

From warren.weckesser at gmail.com  Thu May 26 08:44:07 2016
From: warren.weckesser at gmail.com (Warren Weckesser)
Date: Thu, 26 May 2016 08:44:07 -0400
Subject: [scikit-learn] Fitting Lognormal Distribution
In-Reply-To: <CAPEF3M-JxUD1OUqeuUbiXR6bGTGomFW=gZtARyJxQ5kMCMmSOA@mail.gmail.com>
References: <CAPEF3M-JxUD1OUqeuUbiXR6bGTGomFW=gZtARyJxQ5kMCMmSOA@mail.gmail.com>
Message-ID: <CAGzF1ufjnvSZF+eqyu_PzzGv_4ZjRxHWiQEuGEfAajYuSp=4zw@mail.gmail.com>

On Thu, May 26, 2016 at 2:08 AM, Startup Hire <blrstartuphire at gmail.com>
wrote:

> Hi all,
>
> Hope you are doing good.
>
> I am working on a project where I need to do the following things:
>
> 1. I need to fit a lognormal distribution to a set of values [I know its
> lognormal by a simple XY scatter plot in excel]
>
>

The probability distributions in scipy have a fit() method, and
scipy.stats.lognorm implements the log-normal distribution (
http://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.lognorm.html)
so you can use scipy.lognorm.fit().  See, for example,
http://stackoverflow.com/questions/26406056/a-lognormal-distribution-in-python
or http://stackoverflow.com/
/questions/15630647/fitting-lognormal-distribution-using-scipy-vs-matlab

Warren



> 2. I need to find the intersection of the lognormal distribution so that I
> can decide cut-off values based on that.
>
>
> Can you guide me on (1) and (2) can be achieved in python?
>
> Regards,
> Sanant
>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160526/c0bf7fc7/attachment.html>

From goix.nicolas at gmail.com  Thu May 26 08:50:40 2016
From: goix.nicolas at gmail.com (Nicolas Goix)
Date: Thu, 26 May 2016 14:50:40 +0200
Subject: [scikit-learn] partial_fit implementation for IsolationForest
In-Reply-To: <CAGmDsiMq-s4kwGjVNguESka13dcJGyGs_aZqbNxBBcfFHj3frQ@mail.gmail.com>
References: <e3d4a8743d92e96d3627ce4947271646@cryptolab.net>
 <CAGmDsiMq-s4kwGjVNguESka13dcJGyGs_aZqbNxBBcfFHj3frQ@mail.gmail.com>
Message-ID: <CAPV6P2wkNwNB6aJABqUyqTC8chGmAx22PoVw4ArOedHLSitpSw@mail.gmail.com>

Hello Isaak,

There is a paper from the same authors as iforest but for streaming data:
http://ijcai.org/Proceedings/11/Papers/254.pdf

For now it is not cited enough (24) to satisfy the sklearn requirements.
Waiting for more citations, this could be a nice addition to
sklearn-contrib.

Otherwise, we could imagine extending iforest to streaming data by building
new
trees when data come (and removing the oldest ones), prediction still being
based on
the average depth of the forest. I'm not sure this heuristic could be
merged on
scikit-learn, since it is not based on well-cited papers. In the same time,
it is a natural and simple extension of iforest to streaming data...

Any opinion on it?

Nicolas

2016-05-26 13:32 GMT+02:00 Arthur Mensch <arthur.mensch at inria.fr>:

> Hi Isaac,
>
> You may have a look at MiniBatchKMeans and MiniBatchDictionaryLearning
> that both proposes this API. At the moment, you should fit a single mini
> batch to the estimator using partial_fit, and update the inner attributes
> accordingly. During the first partial_fit, you should take care of various
> memory allocation that are needed by the estimator.
>
> Please fill free to create a pull request whenever you think your code is
> ready for review.
>
> Good luck!
> Le 26 mai 2016 13:14, <donkey-hotei at cryptolab.net> a ?crit :
>
>> hello scikit-learn devs,
>>
>> After following the work on IsolationForest so far and testing on a
>> real-world problem here we've found this model to be very promising for
>> anomaly detection. However, at present, IsolationForest only fits data in
>> batch even while it may be well suited to incremental on-line learning
>> since one could subsample recent history and older estimators can be
>> dropped progressively.
>>
>> I'd like to contribute this feature, but being new to ML and scikit-learn
>> I'm curious how I should start making a quick & dirty version to see how
>> this may work. Are there other good examples where one could see the
>> difference between .fit and .partial_fit in other models?
>>
>> thanks
>> isaak y.
>> _______________________________________________
>> scikit-learn mailing list
>> scikit-learn at python.org
>> https://mail.python.org/mailman/listinfo/scikit-learn
>>
>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160526/72e9681e/attachment.html>

From Dale.T.Smith at macys.com  Thu May 26 09:28:16 2016
From: Dale.T.Smith at macys.com (Dale T Smith)
Date: Thu, 26 May 2016 13:28:16 +0000
Subject: [scikit-learn] partial_fit implementation for IsolationForest
In-Reply-To: <CAPV6P2wkNwNB6aJABqUyqTC8chGmAx22PoVw4ArOedHLSitpSw@mail.gmail.com>
References: <e3d4a8743d92e96d3627ce4947271646@cryptolab.net>
 <CAGmDsiMq-s4kwGjVNguESka13dcJGyGs_aZqbNxBBcfFHj3frQ@mail.gmail.com>
 <CAPV6P2wkNwNB6aJABqUyqTC8chGmAx22PoVw4ArOedHLSitpSw@mail.gmail.com>
Message-ID: <BL2PR06MB227689268E9F94E4E5FDB409C3410@BL2PR06MB2276.namprd06.prod.outlook.com>

I think your idea is an excellent candidate for scikit-learn-contrib

https://github.com/scikit-learn-contrib/scikit-learn-contrib

__________________________________________________________________________________________
Dale Smith | Macy's Systems and Technology | IFS eCommerce | Data Science and Capacity Planning
| 5985 State Bridge Road, Johns Creek, GA 30097 | dale.t.smith at macys.com

From: scikit-learn [mailto:scikit-learn-bounces+dale.t.smith=macys.com at python.org] On Behalf Of Nicolas Goix
Sent: Thursday, May 26, 2016 8:51 AM
To: Scikit-learn user and developer mailing list
Subject: Re: [scikit-learn] partial_fit implementation for IsolationForest

? EXT MSG:
Hello Isaak,

There is a paper from the same authors as iforest but for streaming data: http://ijcai.org/Proceedings/11/Papers/254.pdf

For now it is not cited enough (24) to satisfy the sklearn requirements. Waiting for more citations, this could be a nice addition to sklearn-contrib.

Otherwise, we could imagine extending iforest to streaming data by building new
trees when data come (and removing the oldest ones), prediction still being based on
the average depth of the forest. I'm not sure this heuristic could be merged on
scikit-learn, since it is not based on well-cited papers. In the same time,
it is a natural and simple extension of iforest to streaming data...

Any opinion on it?

Nicolas

2016-05-26 13:32 GMT+02:00 Arthur Mensch <arthur.mensch at inria.fr<mailto:arthur.mensch at inria.fr>>:

Hi Isaac,

You may have a look at MiniBatchKMeans and MiniBatchDictionaryLearning that both proposes this API. At the moment, you should fit a single mini batch to the estimator using partial_fit, and update the inner attributes accordingly. During the first partial_fit, you should take care of various memory allocation that are needed by the estimator.

Please fill free to create a pull request whenever you think your code is ready for review.

Good luck!
Le 26 mai 2016 13:14, <donkey-hotei at cryptolab.net<mailto:donkey-hotei at cryptolab.net>> a ?crit :
hello scikit-learn devs,

After following the work on IsolationForest so far and testing on a real-world problem here we've found this model to be very promising for anomaly detection. However, at present, IsolationForest only fits data in batch even while it may be well suited to incremental on-line learning since one could subsample recent history and older estimators can be dropped progressively.

I'd like to contribute this feature, but being new to ML and scikit-learn I'm curious how I should start making a quick & dirty version to see how this may work. Are there other good examples where one could see the difference between .fit and .partial_fit in other models?

thanks
isaak y.
_______________________________________________
scikit-learn mailing list
scikit-learn at python.org<mailto:scikit-learn at python.org>
https://mail.python.org/mailman/listinfo/scikit-learn

_______________________________________________
scikit-learn mailing list
scikit-learn at python.org<mailto:scikit-learn at python.org>
https://mail.python.org/mailman/listinfo/scikit-learn

* This is an EXTERNAL EMAIL. Stop and think before clicking a link or opening attachments.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160526/8dce1808/attachment-0001.html>

From masad.801 at gmail.com  Thu May 26 09:41:18 2016
From: masad.801 at gmail.com (M Asad)
Date: Thu, 26 May 2016 14:41:18 +0100
Subject: [scikit-learn] Digest Subscription
In-Reply-To: <833806CC-4A4D-4489-B732-0CAA0EC82EA9@gmail.com>
References: <833806CC-4A4D-4489-B732-0CAA0EC82EA9@gmail.com>
Message-ID: <CACNH5NHk9GUC80-oQkcoZ=yrfHCnvPFz=vkfhYmyX-vuWLqUxg@mail.gmail.com>

Hi all,

sorry to bother everyone. I want to remove myself from this list. I cannot
seem to do this via the online portal here:
https://mail.python.org/mailman/listinfo/scikit-learn
When clicking in unsubscribe it does not do anything.

Can anyone please fix this?

On 26 May 2016 at 10:03, Mamun Rashid <mamunbabu2001 at gmail.com> wrote:

> Hi All,
> Since the mailing list switch I am receiving individual emails instead of
> digests. I can not seem to change because for some reason my old password
> is not working.
> Any comments ?
>
> Thanks,
> Mamun
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160526/a6f93ca2/attachment.html>

From sbenthall at gmail.com  Thu May 26 10:20:37 2016
From: sbenthall at gmail.com (Sebastian Benthall)
Date: Thu, 26 May 2016 07:20:37 -0700
Subject: [scikit-learn] Fitting Lognormal Distribution
In-Reply-To: <CAGzF1ufjnvSZF+eqyu_PzzGv_4ZjRxHWiQEuGEfAajYuSp=4zw@mail.gmail.com>
References: <CAPEF3M-JxUD1OUqeuUbiXR6bGTGomFW=gZtARyJxQ5kMCMmSOA@mail.gmail.com>
 <CAGzF1ufjnvSZF+eqyu_PzzGv_4ZjRxHWiQEuGEfAajYuSp=4zw@mail.gmail.com>
Message-ID: <CAEYE9Oe31ROec+HxrWqNwFQFJ3nbZ+9npqxv+eS0rGD7u-pjew@mail.gmail.com>

You may also be interested in the 'powerlaw' Python package, which detects
the tail cutoff.
On May 26, 2016 5:46 AM, "Warren Weckesser" <warren.weckesser at gmail.com>
wrote:

>
>
> On Thu, May 26, 2016 at 2:08 AM, Startup Hire <blrstartuphire at gmail.com>
> wrote:
>
>> Hi all,
>>
>> Hope you are doing good.
>>
>> I am working on a project where I need to do the following things:
>>
>> 1. I need to fit a lognormal distribution to a set of values [I know its
>> lognormal by a simple XY scatter plot in excel]
>>
>>
>
> The probability distributions in scipy have a fit() method, and
> scipy.stats.lognorm implements the log-normal distribution (
> http://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.lognorm.html)
> so you can use scipy.lognorm.fit().  See, for example,
> http://stackoverflow.com/questions/26406056/a-lognormal-distribution-in-python
> or http://stackoverflow.com/
> /questions/15630647/fitting-lognormal-distribution-using-scipy-vs-matlab
>
> Warren
>
>
>
>> 2. I need to find the intersection of the lognormal distribution so that
>> I can decide cut-off values based on that.
>>
>>
>> Can you guide me on (1) and (2) can be achieved in python?
>>
>> Regards,
>> Sanant
>>
>> _______________________________________________
>> scikit-learn mailing list
>> scikit-learn at python.org
>> https://mail.python.org/mailman/listinfo/scikit-learn
>>
>>
>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160526/8b743c81/attachment.html>

From t3kcit at gmail.com  Thu May 26 13:49:47 2016
From: t3kcit at gmail.com (Andreas Mueller)
Date: Thu, 26 May 2016 13:49:47 -0400
Subject: [scikit-learn] Digest Subscription
In-Reply-To: <833806CC-4A4D-4489-B732-0CAA0EC82EA9@gmail.com>
References: <833806CC-4A4D-4489-B732-0CAA0EC82EA9@gmail.com>
Message-ID: <5747373B.2010204@gmail.com>

Hi Mamun.
Sorry my bad.
You can reset your password here:
https://mail.python.org/mailman/listinfo/scikit-learn

Best,
Andy

On 05/26/2016 05:03 AM, Mamun Rashid wrote:
> Hi All,
> Since the mailing list switch I am receiving individual emails instead of digests. I can not seem to change because for some reason my old password is not working.
> Any comments ?
>
> Thanks,
> Mamun
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn


From t3kcit at gmail.com  Thu May 26 13:50:46 2016
From: t3kcit at gmail.com (Andreas Mueller)
Date: Thu, 26 May 2016 13:50:46 -0400
Subject: [scikit-learn] Digest Subscription
In-Reply-To: <CACNH5NHk9GUC80-oQkcoZ=yrfHCnvPFz=vkfhYmyX-vuWLqUxg@mail.gmail.com>
References: <833806CC-4A4D-4489-B732-0CAA0EC82EA9@gmail.com>
 <CACNH5NHk9GUC80-oQkcoZ=yrfHCnvPFz=vkfhYmyX-vuWLqUxg@mail.gmail.com>
Message-ID: <57473776.7030205@gmail.com>

Check your spam folder for a confirmation mail.

On 05/26/2016 09:41 AM, M Asad wrote:
> Hi all,
>
> sorry to bother everyone. I want to remove myself from this list. I 
> cannot seem to do this via the online portal here: 
> https://mail.python.org/mailman/listinfo/scikit-learn
> When clicking in unsubscribe it does not do anything.
>
> Can anyone please fix this?
>
> On 26 May 2016 at 10:03, Mamun Rashid <mamunbabu2001 at gmail.com 
> <mailto:mamunbabu2001 at gmail.com>> wrote:
>
>     Hi All,
>     Since the mailing list switch I am receiving individual emails
>     instead of digests. I can not seem to change because for some
>     reason my old password is not working.
>     Any comments ?
>
>     Thanks,
>     Mamun
>     _______________________________________________
>     scikit-learn mailing list
>     scikit-learn at python.org <mailto:scikit-learn at python.org>
>     https://mail.python.org/mailman/listinfo/scikit-learn
>
>
>
>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160526/b03f2666/attachment.html>

From blrstartuphire at gmail.com  Fri May 27 02:08:01 2016
From: blrstartuphire at gmail.com (Startup Hire)
Date: Fri, 27 May 2016 11:38:01 +0530
Subject: [scikit-learn] Fitting Lognormal Distribution
In-Reply-To: <CAEYE9Oe31ROec+HxrWqNwFQFJ3nbZ+9npqxv+eS0rGD7u-pjew@mail.gmail.com>
References: <CAPEF3M-JxUD1OUqeuUbiXR6bGTGomFW=gZtARyJxQ5kMCMmSOA@mail.gmail.com>
 <CAGzF1ufjnvSZF+eqyu_PzzGv_4ZjRxHWiQEuGEfAajYuSp=4zw@mail.gmail.com>
 <CAEYE9Oe31ROec+HxrWqNwFQFJ3nbZ+9npqxv+eS0rGD7u-pjew@mail.gmail.com>
Message-ID: <CAPEF3M8H5VViJN6FZ1kGw1uJ0hG6dhBxCEarkvBrQ1PN_1W0UA@mail.gmail.com>

Hi,

@ Warren: I was thinking of using federico method as its quite simple. I
know the mu and sigma of log(values) and I need to plot a normal
distribution based on that. Anything inaccurate in doing that?

@ Sebastian: Thanks for your suggestion. I got to know more about powerlaw
distributions.  But, I dont think my values have a long tail. do you think
it is still relevant? What are the potential applications of the same?

Thanks & Regards,
Sanant

On Thu, May 26, 2016 at 7:50 PM, Sebastian Benthall <sbenthall at gmail.com>
wrote:

> You may also be interested in the 'powerlaw' Python package, which detects
> the tail cutoff.
> On May 26, 2016 5:46 AM, "Warren Weckesser" <warren.weckesser at gmail.com>
> wrote:
>
>>
>>
>> On Thu, May 26, 2016 at 2:08 AM, Startup Hire <blrstartuphire at gmail.com>
>> wrote:
>>
>>> Hi all,
>>>
>>> Hope you are doing good.
>>>
>>> I am working on a project where I need to do the following things:
>>>
>>> 1. I need to fit a lognormal distribution to a set of values [I know its
>>> lognormal by a simple XY scatter plot in excel]
>>>
>>>
>>
>> The probability distributions in scipy have a fit() method, and
>> scipy.stats.lognorm implements the log-normal distribution (
>> http://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.lognorm.html)
>> so you can use scipy.lognorm.fit().  See, for example,
>> http://stackoverflow.com/questions/26406056/a-lognormal-distribution-in-python
>> or http://stackoverflow.com/
>> /questions/15630647/fitting-lognormal-distribution-using-scipy-vs-matlab
>>
>> Warren
>>
>>
>>
>>> 2. I need to find the intersection of the lognormal distribution so that
>>> I can decide cut-off values based on that.
>>>
>>>
>>> Can you guide me on (1) and (2) can be achieved in python?
>>>
>>> Regards,
>>> Sanant
>>>
>>> _______________________________________________
>>> scikit-learn mailing list
>>> scikit-learn at python.org
>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>
>>>
>>
>> _______________________________________________
>> scikit-learn mailing list
>> scikit-learn at python.org
>> https://mail.python.org/mailman/listinfo/scikit-learn
>>
>>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160527/fea4fb48/attachment.html>

From warren.weckesser at gmail.com  Fri May 27 09:53:58 2016
From: warren.weckesser at gmail.com (Warren Weckesser)
Date: Fri, 27 May 2016 09:53:58 -0400
Subject: [scikit-learn] Fitting Lognormal Distribution
In-Reply-To: <CAPEF3M8H5VViJN6FZ1kGw1uJ0hG6dhBxCEarkvBrQ1PN_1W0UA@mail.gmail.com>
References: <CAPEF3M-JxUD1OUqeuUbiXR6bGTGomFW=gZtARyJxQ5kMCMmSOA@mail.gmail.com>
 <CAGzF1ufjnvSZF+eqyu_PzzGv_4ZjRxHWiQEuGEfAajYuSp=4zw@mail.gmail.com>
 <CAEYE9Oe31ROec+HxrWqNwFQFJ3nbZ+9npqxv+eS0rGD7u-pjew@mail.gmail.com>
 <CAPEF3M8H5VViJN6FZ1kGw1uJ0hG6dhBxCEarkvBrQ1PN_1W0UA@mail.gmail.com>
Message-ID: <CAGzF1udPXudh-eoOFvjs=NOnHm8whCO9ZzzVaf3pp=63GBv70g@mail.gmail.com>

On Fri, May 27, 2016 at 2:08 AM, Startup Hire <blrstartuphire at gmail.com>
wrote:

> Hi,
>
> @ Warren: I was thinking of using federico method as its quite simple. I
> know the mu and sigma of log(values) and I need to plot a normal
> distribution based on that. Anything inaccurate in doing that?
>
>

Getting mu and sigma from log(values) is fine.  That's one of the three
methods (the one labeled "Explicit formula") that I included in this
answer:
http://stackoverflow.com/questions/15630647/fitting-lognormal-distribution-using-scipy-vs-matlab/15632937#15632937

Warren



> @ Sebastian: Thanks for your suggestion. I got to know more about powerlaw
> distributions.  But, I dont think my values have a long tail. do you think
> it is still relevant? What are the potential applications of the same?
>
> Thanks & Regards,
> Sanant
>
> On Thu, May 26, 2016 at 7:50 PM, Sebastian Benthall <sbenthall at gmail.com>
> wrote:
>
>> You may also be interested in the 'powerlaw' Python package, which
>> detects the tail cutoff.
>> On May 26, 2016 5:46 AM, "Warren Weckesser" <warren.weckesser at gmail.com>
>> wrote:
>>
>>>
>>>
>>> On Thu, May 26, 2016 at 2:08 AM, Startup Hire <blrstartuphire at gmail.com>
>>> wrote:
>>>
>>>> Hi all,
>>>>
>>>> Hope you are doing good.
>>>>
>>>> I am working on a project where I need to do the following things:
>>>>
>>>> 1. I need to fit a lognormal distribution to a set of values [I know
>>>> its lognormal by a simple XY scatter plot in excel]
>>>>
>>>>
>>>
>>> The probability distributions in scipy have a fit() method, and
>>> scipy.stats.lognorm implements the log-normal distribution (
>>> http://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.lognorm.html)
>>> so you can use scipy.lognorm.fit().  See, for example,
>>> http://stackoverflow.com/questions/26406056/a-lognormal-distribution-in-python
>>> or http://stackoverflow.com/
>>> /questions/15630647/fitting-lognormal-distribution-using-scipy-vs-matlab
>>>
>>> Warren
>>>
>>>
>>>
>>>> 2. I need to find the intersection of the lognormal distribution so
>>>> that I can decide cut-off values based on that.
>>>>
>>>>
>>>> Can you guide me on (1) and (2) can be achieved in python?
>>>>
>>>> Regards,
>>>> Sanant
>>>>
>>>> _______________________________________________
>>>> scikit-learn mailing list
>>>> scikit-learn at python.org
>>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>>
>>>>
>>>
>>> _______________________________________________
>>> scikit-learn mailing list
>>> scikit-learn at python.org
>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>
>>>
>> _______________________________________________
>> scikit-learn mailing list
>> scikit-learn at python.org
>> https://mail.python.org/mailman/listinfo/scikit-learn
>>
>>
>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160527/cbe11ffd/attachment-0001.html>

From jmschreiber91 at gmail.com  Fri May 27 12:59:31 2016
From: jmschreiber91 at gmail.com (Jacob Schreiber)
Date: Fri, 27 May 2016 09:59:31 -0700
Subject: [scikit-learn] Fitting Lognormal Distribution
In-Reply-To: <CAGzF1udPXudh-eoOFvjs=NOnHm8whCO9ZzzVaf3pp=63GBv70g@mail.gmail.com>
References: <CAPEF3M-JxUD1OUqeuUbiXR6bGTGomFW=gZtARyJxQ5kMCMmSOA@mail.gmail.com>
 <CAGzF1ufjnvSZF+eqyu_PzzGv_4ZjRxHWiQEuGEfAajYuSp=4zw@mail.gmail.com>
 <CAEYE9Oe31ROec+HxrWqNwFQFJ3nbZ+9npqxv+eS0rGD7u-pjew@mail.gmail.com>
 <CAPEF3M8H5VViJN6FZ1kGw1uJ0hG6dhBxCEarkvBrQ1PN_1W0UA@mail.gmail.com>
 <CAGzF1udPXudh-eoOFvjs=NOnHm8whCO9ZzzVaf3pp=63GBv70g@mail.gmail.com>
Message-ID: <CA+ad8Eu9K1CvzkiZjs1uKJKc-wGMUGBJV95RjtBU7YA-HQ3Quw@mail.gmail.com>

Another option is to use pomegranate
<https://github.com/jmschrei/pomegranate> which has probability
distribution fitting with the same API as scikit-learn. You can see a tutorials
here
<https://github.com/jmschrei/pomegranate/blob/master/tutorials/Tutorial_1_Distributions.ipynb>
and
it includes LogNormalDistribution, in addition to a lot of others. All
distributions also have plotting methods.

On Fri, May 27, 2016 at 6:53 AM, Warren Weckesser <
warren.weckesser at gmail.com> wrote:

>
>
> On Fri, May 27, 2016 at 2:08 AM, Startup Hire <blrstartuphire at gmail.com>
> wrote:
>
>> Hi,
>>
>> @ Warren: I was thinking of using federico method as its quite simple. I
>> know the mu and sigma of log(values) and I need to plot a normal
>> distribution based on that. Anything inaccurate in doing that?
>>
>>
>
> Getting mu and sigma from log(values) is fine.  That's one of the three
> methods (the one labeled "Explicit formula") that I included in this
> answer:
> http://stackoverflow.com/questions/15630647/fitting-lognormal-distribution-using-scipy-vs-matlab/15632937#15632937
>
> Warren
>
>
>
>> @ Sebastian: Thanks for your suggestion. I got to know more about
>> powerlaw distributions.  But, I dont think my values have a long tail. do
>> you think it is still relevant? What are the potential applications of the
>> same?
>>
>> Thanks & Regards,
>> Sanant
>>
>> On Thu, May 26, 2016 at 7:50 PM, Sebastian Benthall <sbenthall at gmail.com>
>> wrote:
>>
>>> You may also be interested in the 'powerlaw' Python package, which
>>> detects the tail cutoff.
>>> On May 26, 2016 5:46 AM, "Warren Weckesser" <warren.weckesser at gmail.com>
>>> wrote:
>>>
>>>>
>>>>
>>>> On Thu, May 26, 2016 at 2:08 AM, Startup Hire <blrstartuphire at gmail.com
>>>> > wrote:
>>>>
>>>>> Hi all,
>>>>>
>>>>> Hope you are doing good.
>>>>>
>>>>> I am working on a project where I need to do the following things:
>>>>>
>>>>> 1. I need to fit a lognormal distribution to a set of values [I know
>>>>> its lognormal by a simple XY scatter plot in excel]
>>>>>
>>>>>
>>>>
>>>> The probability distributions in scipy have a fit() method, and
>>>> scipy.stats.lognorm implements the log-normal distribution (
>>>> http://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.lognorm.html)
>>>> so you can use scipy.lognorm.fit().  See, for example,
>>>> http://stackoverflow.com/questions/26406056/a-lognormal-distribution-in-python
>>>> or http://stackoverflow.com/
>>>> /questions/15630647/fitting-lognormal-distribution-using-scipy-vs-matlab
>>>>
>>>> Warren
>>>>
>>>>
>>>>
>>>>> 2. I need to find the intersection of the lognormal distribution so
>>>>> that I can decide cut-off values based on that.
>>>>>
>>>>>
>>>>> Can you guide me on (1) and (2) can be achieved in python?
>>>>>
>>>>> Regards,
>>>>> Sanant
>>>>>
>>>>> _______________________________________________
>>>>> scikit-learn mailing list
>>>>> scikit-learn at python.org
>>>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>>>
>>>>>
>>>>
>>>> _______________________________________________
>>>> scikit-learn mailing list
>>>> scikit-learn at python.org
>>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>>
>>>>
>>> _______________________________________________
>>> scikit-learn mailing list
>>> scikit-learn at python.org
>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>
>>>
>>
>> _______________________________________________
>> scikit-learn mailing list
>> scikit-learn at python.org
>> https://mail.python.org/mailman/listinfo/scikit-learn
>>
>>
>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160527/cb066194/attachment.html>

From t3kcit at gmail.com  Fri May 27 18:10:41 2016
From: t3kcit at gmail.com (Andreas Mueller)
Date: Fri, 27 May 2016 18:10:41 -0400
Subject: [scikit-learn] partial_fit implementation for IsolationForest
In-Reply-To: <BL2PR06MB227689268E9F94E4E5FDB409C3410@BL2PR06MB2276.namprd06.prod.outlook.com>
References: <e3d4a8743d92e96d3627ce4947271646@cryptolab.net>
 <CAGmDsiMq-s4kwGjVNguESka13dcJGyGs_aZqbNxBBcfFHj3frQ@mail.gmail.com>
 <CAPV6P2wkNwNB6aJABqUyqTC8chGmAx22PoVw4ArOedHLSitpSw@mail.gmail.com>
 <BL2PR06MB227689268E9F94E4E5FDB409C3410@BL2PR06MB2276.namprd06.prod.outlook.com>
Message-ID: <5748C5E1.4090303@gmail.com>

How about mondrian forests ;)


On 05/26/2016 09:28 AM, Dale T Smith wrote:
>
> I think your idea is an excellent candidate for scikit-learn-contrib
>
> https://github.com/scikit-learn-contrib/scikit-learn-contrib
>
> __________________________________________________________________________________________
> *Dale Smith*| Macy's Systems and Technology | IFS eCommerce | Data 
> Science and Capacity Planning
> | 5985 State Bridge Road, Johns Creek, GA 30097 | dale.t.smith at macys.com
>
> *From:*scikit-learn 
> [mailto:scikit-learn-bounces+dale.t.smith=macys.com at python.org] *On 
> Behalf Of *Nicolas Goix
> *Sent:* Thursday, May 26, 2016 8:51 AM
> *To:* Scikit-learn user and developer mailing list
> *Subject:* Re: [scikit-learn] partial_fit implementation for 
> IsolationForest
>
> ? EXT MSG:
>
> Hello Isaak,
>
> There is a paper from the same authors as iforest but for streaming 
> data: http://ijcai.org/Proceedings/11/Papers/254.pdf
>
>
> For now it is not cited enough (24) to satisfy the sklearn 
> requirements. Waiting for more citations, this could be a nice 
> addition to sklearn-contrib.
>
> Otherwise, we could imagine extending iforest to streaming data by 
> building new
> trees when data come (and removing the oldest ones), prediction still 
> being based on
> the average depth of the forest. I'm not sure this heuristic could be 
> merged on
> scikit-learn, since it is not based on well-cited papers. In the same 
> time,
> it is a natural and simple extension of iforest to streaming data...
>
> Any opinion on it?
>
> Nicolas
>
> 2016-05-26 13:32 GMT+02:00 Arthur Mensch <arthur.mensch at inria.fr 
> <mailto:arthur.mensch at inria.fr>>:
>
> Hi Isaac,
>
> You may have a look at MiniBatchKMeans and MiniBatchDictionaryLearning 
> that both proposes this API. At the moment, you should fit a single 
> mini batch to the estimator using partial_fit, and update the inner 
> attributes accordingly. During the first partial_fit, you should take 
> care of various memory allocation that are needed by the estimator.
>
> Please fill free to create a pull request whenever you think your code 
> is ready for review.
>
> Good luck!
>
> Le 26 mai 2016 13:14, <donkey-hotei at cryptolab.net 
> <mailto:donkey-hotei at cryptolab.net>> a ?crit :
>
> hello scikit-learn devs,
>
> After following the work on IsolationForest so far and testing on a 
> real-world problem here we've found this model to be very promising 
> for anomaly detection. However, at present, IsolationForest only fits 
> data in batch even while it may be well suited to incremental on-line 
> learning since one could subsample recent history and older estimators 
> can be dropped progressively.
>
> I'd like to contribute this feature, but being new to ML and 
> scikit-learn I'm curious how I should start making a quick & dirty 
> version to see how this may work. Are there other good examples where 
> one could see the difference between .fit and .partial_fit in other 
> models?
>
> thanks
> isaak y.
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org <mailto:scikit-learn at python.org>
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org <mailto:scikit-learn at python.org>
> https://mail.python.org/mailman/listinfo/scikit-learn
>
> * This is an EXTERNAL EMAIL. Stop and think before clicking a link or 
> opening attachments.
>
>
>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160527/5fba0d2c/attachment-0001.html>

From roberto.pagliari at asos.com  Sun May 29 09:48:13 2016
From: roberto.pagliari at asos.com (Roberto Pagliari)
Date: Sun, 29 May 2016 13:48:13 +0000
Subject: [scikit-learn] joblib deprecation warning
Message-ID: <F5C85AA8-AB99-47B1-8C99-45DA18F9AFB9@asos.com>

I'm getting the following warning with anaconda


anaconda2/lib/python2.7/site-packages/sklearn/externals/joblib/hashing.py:197: DeprecationWarning: Changing the shape of non-C contiguous array by

descriptor assignment is deprecated. To maintain

the Fortran contiguity of a multidimensional Fortran

array, use 'a.T.view(...).T' instead

  obj_bytes_view = obj.view(self.np.uint8)

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160529/49182893/attachment.html>

From roberto.pagliari at asos.com  Sun May 29 10:03:21 2016
From: roberto.pagliari at asos.com (Roberto Pagliari)
Date: Sun, 29 May 2016 14:03:21 +0000
Subject: [scikit-learn] random forest regression with multiple output
 variables
Message-ID: <DDD4E01E-525C-4CE3-91A0-DCA07D25677F@asos.com>

When using random forest regressor, the default scoring function is mse.

In the case of multiple output, is the mse the sum over all output variables?

If so, are the output variables scaled, to make the mse comparable across different variables?

Thank you,

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160529/45a6a0fb/attachment.html>

From alexandre.abadie at inria.fr  Sun May 29 11:58:26 2016
From: alexandre.abadie at inria.fr (Alexandre Abadie)
Date: Sun, 29 May 2016 17:58:26 +0200 (CEST)
Subject: [scikit-learn] joblib deprecation warning
In-Reply-To: <F5C85AA8-AB99-47B1-8C99-45DA18F9AFB9@asos.com>
References: <F5C85AA8-AB99-47B1-8C99-45DA18F9AFB9@asos.com>
Message-ID: <1149752649.14254376.1464537506584.JavaMail.zimbra@inria.fr>

Hi,

Thanks for reporting.
This is a known issue from joblib already reported in [1] and fixed by [2].

Alex

[1] https://github.com/joblib/joblib/issues/308
[2] https://github.com/joblib/joblib/pull/309

----- Mail original -----
> I'm getting the following warning with anaconda
> 
> 
> anaconda2/lib/python2.7/site-packages/sklearn/externals/joblib/hashing.py:197:
> DeprecationWarning: Changing the shape of non-C contiguous array by
> 
> descriptor assignment is deprecated. To maintain
> 
> the Fortran contiguity of a multidimensional Fortran
> 
> array, use 'a.T.view(...).T' instead
> 
>   obj_bytes_view = obj.view(self.np.uint8)
> 
> 
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
> 

From roberto.pagliari at asos.com  Sun May 29 21:34:38 2016
From: roberto.pagliari at asos.com (Roberto Pagliari)
Date: Mon, 30 May 2016 01:34:38 +0000
Subject: [scikit-learn] joblib deprecation warning
Message-ID: <402024B2-AC94-45B6-94A9-FD1A7D5448AB@asos.com>

Thank you, 





On 29/05/2016, 16:58, "scikit-learn on behalf of Alexandre Abadie" <scikit-learn-bounces+roberto.pagliari=asos.com at python.org on behalf of alexandre.abadie at inria.fr> wrote:

>Hi,
>
>Thanks for reporting.
>This is a known issue from joblib already reported in [1] and fixed by [2].
>
>Alex
>
>[1] https://github.com/joblib/joblib/issues/308
>[2] https://github.com/joblib/joblib/pull/309
>
>----- Mail original -----
>> I'm getting the following warning with anaconda
>> 
>> 
>> anaconda2/lib/python2.7/site-packages/sklearn/externals/joblib/hashing.py:197:
>> DeprecationWarning: Changing the shape of non-C contiguous array by
>> 
>> descriptor assignment is deprecated. To maintain
>> 
>> the Fortran contiguity of a multidimensional Fortran
>> 
>> array, use 'a.T.view(...).T' instead
>> 
>>   obj_bytes_view = obj.view(self.np.uint8)
>> 
>> 
>> _______________________________________________
>> scikit-learn mailing list
>> scikit-learn at python.org
>> https://mail.python.org/mailman/listinfo/scikit-learn
>> 
>_______________________________________________
>scikit-learn mailing list
>scikit-learn at python.org
>https://mail.python.org/mailman/listinfo/scikit-learn

From blrstartuphire at gmail.com  Mon May 30 07:18:57 2016
From: blrstartuphire at gmail.com (Startup Hire)
Date: Mon, 30 May 2016 16:48:57 +0530
Subject: [scikit-learn] Fitting Lognormal Distribution
In-Reply-To: <CA+ad8Eu9K1CvzkiZjs1uKJKc-wGMUGBJV95RjtBU7YA-HQ3Quw@mail.gmail.com>
References: <CAPEF3M-JxUD1OUqeuUbiXR6bGTGomFW=gZtARyJxQ5kMCMmSOA@mail.gmail.com>
 <CAGzF1ufjnvSZF+eqyu_PzzGv_4ZjRxHWiQEuGEfAajYuSp=4zw@mail.gmail.com>
 <CAEYE9Oe31ROec+HxrWqNwFQFJ3nbZ+9npqxv+eS0rGD7u-pjew@mail.gmail.com>
 <CAPEF3M8H5VViJN6FZ1kGw1uJ0hG6dhBxCEarkvBrQ1PN_1W0UA@mail.gmail.com>
 <CAGzF1udPXudh-eoOFvjs=NOnHm8whCO9ZzzVaf3pp=63GBv70g@mail.gmail.com>
 <CA+ad8Eu9K1CvzkiZjs1uKJKc-wGMUGBJV95RjtBU7YA-HQ3Quw@mail.gmail.com>
Message-ID: <CAPEF3M8W+j+yHKO7=eP1vsMQojN+k_rX6rM+=UtDGOSJms0DWQ@mail.gmail.com>

Thanks to all the replies.

I was able to write the intial code

- Refer the charts below.. After the second red point, can I say that the
values of "BLUE" curve will always be higher than "GREEN" curve?

   -  The ultimate objective is to find out when the values of blue curve
   starts exceeding the values of green curve.





 Regards, Sanant[image: Inline image 1]

On Fri, May 27, 2016 at 10:29 PM, Jacob Schreiber <jmschreiber91 at gmail.com>
wrote:

> Another option is to use pomegranate
> <https://github.com/jmschrei/pomegranate> which has probability
> distribution fitting with the same API as scikit-learn. You can see a tutorials
> here
> <https://github.com/jmschrei/pomegranate/blob/master/tutorials/Tutorial_1_Distributions.ipynb> and
> it includes LogNormalDistribution, in addition to a lot of others. All
> distributions also have plotting methods.
>
> On Fri, May 27, 2016 at 6:53 AM, Warren Weckesser <
> warren.weckesser at gmail.com> wrote:
>
>>
>>
>> On Fri, May 27, 2016 at 2:08 AM, Startup Hire <blrstartuphire at gmail.com>
>> wrote:
>>
>>> Hi,
>>>
>>> @ Warren: I was thinking of using federico method as its quite simple. I
>>> know the mu and sigma of log(values) and I need to plot a normal
>>> distribution based on that. Anything inaccurate in doing that?
>>>
>>>
>>
>> Getting mu and sigma from log(values) is fine.  That's one of the three
>> methods (the one labeled "Explicit formula") that I included in this
>> answer:
>> http://stackoverflow.com/questions/15630647/fitting-lognormal-distribution-using-scipy-vs-matlab/15632937#15632937
>>
>> Warren
>>
>>
>>
>>> @ Sebastian: Thanks for your suggestion. I got to know more about
>>> powerlaw distributions.  But, I dont think my values have a long tail. do
>>> you think it is still relevant? What are the potential applications of the
>>> same?
>>>
>>> Thanks & Regards,
>>> Sanant
>>>
>>> On Thu, May 26, 2016 at 7:50 PM, Sebastian Benthall <sbenthall at gmail.com
>>> > wrote:
>>>
>>>> You may also be interested in the 'powerlaw' Python package, which
>>>> detects the tail cutoff.
>>>> On May 26, 2016 5:46 AM, "Warren Weckesser" <warren.weckesser at gmail.com>
>>>> wrote:
>>>>
>>>>>
>>>>>
>>>>> On Thu, May 26, 2016 at 2:08 AM, Startup Hire <
>>>>> blrstartuphire at gmail.com> wrote:
>>>>>
>>>>>> Hi all,
>>>>>>
>>>>>> Hope you are doing good.
>>>>>>
>>>>>> I am working on a project where I need to do the following things:
>>>>>>
>>>>>> 1. I need to fit a lognormal distribution to a set of values [I know
>>>>>> its lognormal by a simple XY scatter plot in excel]
>>>>>>
>>>>>>
>>>>>
>>>>> The probability distributions in scipy have a fit() method, and
>>>>> scipy.stats.lognorm implements the log-normal distribution (
>>>>> http://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.lognorm.html)
>>>>> so you can use scipy.lognorm.fit().  See, for example,
>>>>> http://stackoverflow.com/questions/26406056/a-lognormal-distribution-in-python
>>>>> or http://stackoverflow.com/
>>>>>
>>>>> /questions/15630647/fitting-lognormal-distribution-using-scipy-vs-matlab
>>>>>
>>>>> Warren
>>>>>
>>>>>
>>>>>
>>>>>> 2. I need to find the intersection of the lognormal distribution so
>>>>>> that I can decide cut-off values based on that.
>>>>>>
>>>>>>
>>>>>> Can you guide me on (1) and (2) can be achieved in python?
>>>>>>
>>>>>> Regards,
>>>>>> Sanant
>>>>>>
>>>>>> _______________________________________________
>>>>>> scikit-learn mailing list
>>>>>> scikit-learn at python.org
>>>>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>>>>
>>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> scikit-learn mailing list
>>>>> scikit-learn at python.org
>>>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>>>
>>>>>
>>>> _______________________________________________
>>>> scikit-learn mailing list
>>>> scikit-learn at python.org
>>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>>
>>>>
>>>
>>> _______________________________________________
>>> scikit-learn mailing list
>>> scikit-learn at python.org
>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>
>>>
>>
>> _______________________________________________
>> scikit-learn mailing list
>> scikit-learn at python.org
>> https://mail.python.org/mailman/listinfo/scikit-learn
>>
>>
>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160530/4ce99bc0/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Curve 1 and Curve 2.png
Type: image/png
Size: 23093 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160530/4ce99bc0/attachment-0001.png>

From roberto.pagliari at asos.com  Mon May 30 10:05:40 2016
From: roberto.pagliari at asos.com (Roberto Pagliari)
Date: Mon, 30 May 2016 14:05:40 +0000
Subject: [scikit-learn] gradient boosting regression with multiple output
 variables?
Message-ID: <63E8D7E6-D4A4-443E-ADA9-4C1213F7774F@asos.com>

I noticed that the fit method of GBR does not return a [n_samples, n_output] array. Does that mean multiple output variables are not supported?

I'm asking because most other regressors do.

Thank you,

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160530/dc09c2ee/attachment.html>

From peter.prettenhofer at gmail.com  Mon May 30 12:09:39 2016
From: peter.prettenhofer at gmail.com (Peter Prettenhofer)
Date: Mon, 30 May 2016 18:09:39 +0200
Subject: [scikit-learn] gradient boosting regression with multiple
 output variables?
In-Reply-To: <63E8D7E6-D4A4-443E-ADA9-4C1213F7774F@asos.com>
References: <63E8D7E6-D4A4-443E-ADA9-4C1213F7774F@asos.com>
Message-ID: <CAD-7WjoDHZtX16HMUkQsWSXLWG5Urz8cDOywP9_rju3Fjt++=g@mail.gmail.com>

Hi Roberto,

correct - GradientBoostingRegressor | GradientBoostingClassifier does not
support multiple outputs.

best,
 Peter

2016-05-30 16:05 GMT+02:00 Roberto Pagliari <roberto.pagliari at asos.com>:

> I noticed that the fit method of GBR does not return a [n_samples,
> n_output] array. Does that mean multiple output variables are not supported?
>
> I'm asking because most other regressors do.
>
> Thank you,
>
>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>


-- 
Peter Prettenhofer
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160530/cc429fe8/attachment.html>

From betatim at gmail.com  Mon May 30 12:56:03 2016
From: betatim at gmail.com (Tim Head)
Date: Mon, 30 May 2016 16:56:03 +0000
Subject: [scikit-learn] gradient boosting regression with multiple
 output variables?
In-Reply-To: <CAD-7WjoDHZtX16HMUkQsWSXLWG5Urz8cDOywP9_rju3Fjt++=g@mail.gmail.com>
References: <63E8D7E6-D4A4-443E-ADA9-4C1213F7774F@asos.com>
 <CAD-7WjoDHZtX16HMUkQsWSXLWG5Urz8cDOywP9_rju3Fjt++=g@mail.gmail.com>
Message-ID: <CAN3x1RYtW=gDJip+QoPfeNCM1Kj-ev4cQhL-vwiPrw871FPrag@mail.gmail.com>

Hi,

recently a MultiOutput* "adaptor" was added to scikit-learn (-dev version
only so far I think). They take algorithms like GradientBoosting* and fit
one instance per target wrapped in a nice interface. You won't be able to
take advantage of correlations between the outputs this way but it might be
a starting point. Take a look at:

http://scikit-learn.org/dev/modules/classes.html#module-sklearn.multioutput
http://scikit-learn.org/dev/auto_examples/ensemble/plot_random_forest_regression_multioutput.html

T

On Mon, May 30, 2016 at 6:16 PM Peter Prettenhofer <
peter.prettenhofer at gmail.com> wrote:

> Hi Roberto,
>
> correct - GradientBoostingRegressor | GradientBoostingClassifier does not
> support multiple outputs.
>
> best,
>  Peter
>
> 2016-05-30 16:05 GMT+02:00 Roberto Pagliari <roberto.pagliari at asos.com>:
>
>> I noticed that the fit method of GBR does not return a [n_samples,
>> n_output] array. Does that mean multiple output variables are not supported?
>>
>> I'm asking because most other regressors do.
>>
>> Thank you,
>>
>>
>> _______________________________________________
>> scikit-learn mailing list
>> scikit-learn at python.org
>> https://mail.python.org/mailman/listinfo/scikit-learn
>>
>>
>
>
> --
> Peter Prettenhofer
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160530/c146e9b3/attachment.html>

From arnaud4567 at gmail.com  Mon May 30 17:55:59 2016
From: arnaud4567 at gmail.com (Arnaud Joly)
Date: Mon, 30 May 2016 23:55:59 +0200
Subject: [scikit-learn] random forest regression with multiple output
 variables
In-Reply-To: <DDD4E01E-525C-4CE3-91A0-DCA07D25677F@asos.com>
References: <DDD4E01E-525C-4CE3-91A0-DCA07D25677F@asos.com>
Message-ID: <B54908E8-8DD0-4BE9-B374-58A9FEECA6CA@gmail.com>

Hi,


> On 29 May 2016, at 16:03, Roberto Pagliari <roberto.pagliari at asos.com> wrote:
> 
> When using random forest regressor, the default scoring function is mse. 
> 
> In the case of multiple output, is the mse the sum over all output variables?
> 

Yes, it is summed over all outputs.


> If so, are the output variables scaled, to make the mse comparable across different variables?
> 

No, they aren?t scaled.

Best regards,
Arnaud Joly
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160530/30c4eacb/attachment.html>

From m.waseem.ahmad at gmail.com  Tue May 31 11:55:10 2016
From: m.waseem.ahmad at gmail.com (muhammad waseem)
Date: Tue, 31 May 2016 16:55:10 +0100
Subject: [scikit-learn] Artificial neural network not learning lower values
 of the training sample
Message-ID: <CAGWVJJWUJCFbyA7iKd8dK3qQe5nC2xu1Hm_ox_S-XYPKNkt5-Q@mail.gmail.com>

Hi All,
I am trying to train an ANN but until now it is not learning the lower
values of the training sample. I have tried using different python
libraries to train ANN. The aim is to predict solar radiation from other
weather parameters (regression problem). I think the ANN is confusing lower
values (winter/cloudy days) with the night-time values (probably). I have
tried the following but none of them worked;

 1. Scaling data between different values e.g. [0,1],[-1,1]
 2. Standardising data to have zero mean and unit variance
 3. Shuffling the data
 4. Increasing the training samples (from 3 years to 10 years)
 5. Using different train functions
 6. Trying different transfer functions
 7. Using few input variables
 8. Varying hidden layers and hidden layers' neurons

Any idea what could be wrong or any directions to try?

Thanks
Kindest Regards
Waseem
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160531/4c191ba8/attachment.html>

From andrew.matte at gmail.com  Tue May 31 12:42:11 2016
From: andrew.matte at gmail.com (Andrew Matte)
Date: Tue, 31 May 2016 12:42:11 -0400
Subject: [scikit-learn] Artificial neural network not learning lower
 values of the training sample
In-Reply-To: <CANiXb6W4YhQaRGShnAfgiGvK9q-gZ7JGvfwhxVyuapyHPNLuWQ@mail.gmail.com>
References: <CAGWVJJWUJCFbyA7iKd8dK3qQe5nC2xu1Hm_ox_S-XYPKNkt5-Q@mail.gmail.com>
 <CANiXb6VFXYOk8ZiKF2b0BhXTXdaKO+DWnhWb95PbOpCdCxo3AQ@mail.gmail.com>
 <CANiXb6W4YhQaRGShnAfgiGvK9q-gZ7JGvfwhxVyuapyHPNLuWQ@mail.gmail.com>
Message-ID: <CANiXb6V9uKcEnjtwNMZPxpo=U7hWKT1v67AH3cUHSfAWSmJ1nA@mail.gmail.com>

It sounds like you're right about the low values being connected between
night and clouds. Since there's no seasonality outside of recurrent NNs,
maybe try adding an hour of day variable.
On May 31, 2016 11:57 AM, "muhammad waseem" <m.waseem.ahmad at gmail.com>
wrote:

Hi All,
I am trying to train an ANN but until now it is not learning the lower
values of the training sample. I have tried using different python
libraries to train ANN. The aim is to predict solar radiation from other
weather parameters (regression problem). I think the ANN is confusing lower
values (winter/cloudy days) with the night-time values (probably). I have
tried the following but none of them worked;

 1. Scaling data between different values e.g. [0,1],[-1,1]
 2. Standardising data to have zero mean and unit variance
 3. Shuffling the data
 4. Increasing the training samples (from 3 years to 10 years)
 5. Using different train functions
 6. Trying different transfer functions
 7. Using few input variables
 8. Varying hidden layers and hidden layers' neurons

Any idea what could be wrong or any directions to try?

Thanks
Kindest Regards
Waseem

_______________________________________________
scikit-learn mailing list
scikit-learn at python.org
https://mail.python.org/mailman/listinfo/scikit-learn
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160531/8e4705e9/attachment.html>

From andrewholmes82 at icloud.com  Tue May 31 11:59:03 2016
From: andrewholmes82 at icloud.com (Andrew Holmes)
Date: Tue, 31 May 2016 16:59:03 +0100
Subject: [scikit-learn] Artificial neural network not learning lower
 values of the training sample
In-Reply-To: <CAGWVJJWUJCFbyA7iKd8dK3qQe5nC2xu1Hm_ox_S-XYPKNkt5-Q@mail.gmail.com>
References: <CAGWVJJWUJCFbyA7iKd8dK3qQe5nC2xu1Hm_ox_S-XYPKNkt5-Q@mail.gmail.com>
Message-ID: <3FB2A8F0-9EAE-48CB-A0AC-98D9F27EA5A8@icloud.com>

If the problem is that it?s confusing day and night, are you including time of day as a parameter?

Best wishes
Andrew

@andrewholmes82 <http://twitter.com/andrewholmes82>








> On 31 May 2016, at 16:55, muhammad waseem <m.waseem.ahmad at gmail.com> wrote:
> 
> Hi All,
> I am trying to train an ANN but until now it is not learning the lower values of the training sample. I have tried using different python libraries to train ANN. The aim is to predict solar radiation from other weather parameters (regression problem). I think the ANN is confusing lower values (winter/cloudy days) with the night-time values (probably). I have tried the following but none of them worked;
> 
>  1. Scaling data between different values e.g. [0,1],[-1,1]
>  2. Standardising data to have zero mean and unit variance
>  3. Shuffling the data
>  4. Increasing the training samples (from 3 years to 10 years)
>  5. Using different train functions
>  6. Trying different transfer functions
>  7. Using few input variables
>  8. Varying hidden layers and hidden layers' neurons
> 
> Any idea what could be wrong or any directions to try?
> 
> Thanks
> Kindest Regards
> Waseem
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160531/90f8ed8a/attachment.html>

From m.waseem.ahmad at gmail.com  Tue May 31 13:47:27 2016
From: m.waseem.ahmad at gmail.com (muhammad waseem)
Date: Tue, 31 May 2016 18:47:27 +0100
Subject: [scikit-learn] Artificial neural network not learning lower
 values of the training sample
In-Reply-To: <3FB2A8F0-9EAE-48CB-A0AC-98D9F27EA5A8@icloud.com>
References: <CAGWVJJWUJCFbyA7iKd8dK3qQe5nC2xu1Hm_ox_S-XYPKNkt5-Q@mail.gmail.com>
 <3FB2A8F0-9EAE-48CB-A0AC-98D9F27EA5A8@icloud.com>
Message-ID: <CAGWVJJX47yEJF1g7DgSSwX5ChwBH6SXouDB5duo8yZMaBqeRfA@mail.gmail.com>

Thanks for your reply. I have day, month, hour, temp, relative humidity,
Wind speed as my input variables. I can't think of any other dependant
variables. It is quite strange to me that I don't get results after using
these input variables.

On Tue, May 31, 2016 at 4:59 PM, Andrew Holmes <andrewholmes82 at icloud.com>
wrote:

> If the problem is that it?s confusing day and night, are you including
> time of day as a parameter?
>
> Best wishes
> Andrew
>
> @andrewholmes82 <http://twitter.com/andrewholmes82>
>
>
>
>
>
>
>
>
> On 31 May 2016, at 16:55, muhammad waseem <m.waseem.ahmad at gmail.com>
> wrote:
>
> Hi All,
> I am trying to train an ANN but until now it is not learning the lower
> values of the training sample. I have tried using different python
> libraries to train ANN. The aim is to predict solar radiation from other
> weather parameters (regression problem). I think the ANN is confusing lower
> values (winter/cloudy days) with the night-time values (probably). I have
> tried the following but none of them worked;
>
>  1. Scaling data between different values e.g. [0,1],[-1,1]
>  2. Standardising data to have zero mean and unit variance
>  3. Shuffling the data
>  4. Increasing the training samples (from 3 years to 10 years)
>  5. Using different train functions
>  6. Trying different transfer functions
>  7. Using few input variables
>  8. Varying hidden layers and hidden layers' neurons
>
> Any idea what could be wrong or any directions to try?
>
> Thanks
> Kindest Regards
> Waseem
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>
>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160531/4f6bef65/attachment-0001.html>

From jmschreiber91 at gmail.com  Tue May 31 14:07:21 2016
From: jmschreiber91 at gmail.com (Jacob Schreiber)
Date: Tue, 31 May 2016 11:07:21 -0700
Subject: [scikit-learn] Artificial neural network not learning lower
 values of the training sample
In-Reply-To: <CAGWVJJX47yEJF1g7DgSSwX5ChwBH6SXouDB5duo8yZMaBqeRfA@mail.gmail.com>
References: <CAGWVJJWUJCFbyA7iKd8dK3qQe5nC2xu1Hm_ox_S-XYPKNkt5-Q@mail.gmail.com>
 <3FB2A8F0-9EAE-48CB-A0AC-98D9F27EA5A8@icloud.com>
 <CAGWVJJX47yEJF1g7DgSSwX5ChwBH6SXouDB5duo8yZMaBqeRfA@mail.gmail.com>
Message-ID: <CA+ad8EuEb7fLAWgL+ne3D0n+41u_sV-NMFZtNtAzt-LqgRy73w@mail.gmail.com>

Do you have any other baselines which you can compare to? It might be
helpful in seeing if this is a problem which can be learned.

On Tue, May 31, 2016 at 10:47 AM, muhammad waseem <m.waseem.ahmad at gmail.com>
wrote:

> Thanks for your reply. I have day, month, hour, temp, relative humidity,
> Wind speed as my input variables. I can't think of any other dependant
> variables. It is quite strange to me that I don't get results after using
> these input variables.
>
> On Tue, May 31, 2016 at 4:59 PM, Andrew Holmes <andrewholmes82 at icloud.com>
> wrote:
>
>> If the problem is that it?s confusing day and night, are you including
>> time of day as a parameter?
>>
>> Best wishes
>> Andrew
>>
>> @andrewholmes82 <http://twitter.com/andrewholmes82>
>>
>>
>>
>>
>>
>>
>>
>>
>> On 31 May 2016, at 16:55, muhammad waseem <m.waseem.ahmad at gmail.com>
>> wrote:
>>
>> Hi All,
>> I am trying to train an ANN but until now it is not learning the lower
>> values of the training sample. I have tried using different python
>> libraries to train ANN. The aim is to predict solar radiation from other
>> weather parameters (regression problem). I think the ANN is confusing lower
>> values (winter/cloudy days) with the night-time values (probably). I have
>> tried the following but none of them worked;
>>
>>  1. Scaling data between different values e.g. [0,1],[-1,1]
>>  2. Standardising data to have zero mean and unit variance
>>  3. Shuffling the data
>>  4. Increasing the training samples (from 3 years to 10 years)
>>  5. Using different train functions
>>  6. Trying different transfer functions
>>  7. Using few input variables
>>  8. Varying hidden layers and hidden layers' neurons
>>
>> Any idea what could be wrong or any directions to try?
>>
>> Thanks
>> Kindest Regards
>> Waseem
>> _______________________________________________
>> scikit-learn mailing list
>> scikit-learn at python.org
>> https://mail.python.org/mailman/listinfo/scikit-learn
>>
>>
>>
>> _______________________________________________
>> scikit-learn mailing list
>> scikit-learn at python.org
>> https://mail.python.org/mailman/listinfo/scikit-learn
>>
>>
>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160531/5e2b0325/attachment.html>

From m.waseem.ahmad at gmail.com  Tue May 31 14:10:54 2016
From: m.waseem.ahmad at gmail.com (muhammad waseem)
Date: Tue, 31 May 2016 19:10:54 +0100
Subject: [scikit-learn] Artificial neural network not learning lower
 values of the training sample
In-Reply-To: <CA+ad8EuEb7fLAWgL+ne3D0n+41u_sV-NMFZtNtAzt-LqgRy73w@mail.gmail.com>
References: <CAGWVJJWUJCFbyA7iKd8dK3qQe5nC2xu1Hm_ox_S-XYPKNkt5-Q@mail.gmail.com>
 <3FB2A8F0-9EAE-48CB-A0AC-98D9F27EA5A8@icloud.com>
 <CAGWVJJX47yEJF1g7DgSSwX5ChwBH6SXouDB5duo8yZMaBqeRfA@mail.gmail.com>
 <CA+ad8EuEb7fLAWgL+ne3D0n+41u_sV-NMFZtNtAzt-LqgRy73w@mail.gmail.com>
Message-ID: <CAGWVJJXtc2BTRuiPb=noPYXLhttNe0is+U4iZvaduiTCyZGu8g@mail.gmail.com>

This problem has been solved in the literature before, I can post papers.

On Tue, May 31, 2016 at 7:07 PM, Jacob Schreiber <jmschreiber91 at gmail.com>
wrote:

> Do you have any other baselines which you can compare to? It might be
> helpful in seeing if this is a problem which can be learned.
>
> On Tue, May 31, 2016 at 10:47 AM, muhammad waseem <
> m.waseem.ahmad at gmail.com> wrote:
>
>> Thanks for your reply. I have day, month, hour, temp, relative humidity,
>> Wind speed as my input variables. I can't think of any other dependant
>> variables. It is quite strange to me that I don't get results after using
>> these input variables.
>>
>> On Tue, May 31, 2016 at 4:59 PM, Andrew Holmes <andrewholmes82 at icloud.com
>> > wrote:
>>
>>> If the problem is that it?s confusing day and night, are you including
>>> time of day as a parameter?
>>>
>>> Best wishes
>>> Andrew
>>>
>>> @andrewholmes82 <http://twitter.com/andrewholmes82>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>> On 31 May 2016, at 16:55, muhammad waseem <m.waseem.ahmad at gmail.com>
>>> wrote:
>>>
>>> Hi All,
>>> I am trying to train an ANN but until now it is not learning the lower
>>> values of the training sample. I have tried using different python
>>> libraries to train ANN. The aim is to predict solar radiation from other
>>> weather parameters (regression problem). I think the ANN is confusing lower
>>> values (winter/cloudy days) with the night-time values (probably). I have
>>> tried the following but none of them worked;
>>>
>>>  1. Scaling data between different values e.g. [0,1],[-1,1]
>>>  2. Standardising data to have zero mean and unit variance
>>>  3. Shuffling the data
>>>  4. Increasing the training samples (from 3 years to 10 years)
>>>  5. Using different train functions
>>>  6. Trying different transfer functions
>>>  7. Using few input variables
>>>  8. Varying hidden layers and hidden layers' neurons
>>>
>>> Any idea what could be wrong or any directions to try?
>>>
>>> Thanks
>>> Kindest Regards
>>> Waseem
>>> _______________________________________________
>>> scikit-learn mailing list
>>> scikit-learn at python.org
>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>
>>>
>>>
>>> _______________________________________________
>>> scikit-learn mailing list
>>> scikit-learn at python.org
>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>
>>>
>>
>> _______________________________________________
>> scikit-learn mailing list
>> scikit-learn at python.org
>> https://mail.python.org/mailman/listinfo/scikit-learn
>>
>>
>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160531/74f1ffc3/attachment.html>

From jmschreiber91 at gmail.com  Tue May 31 14:18:23 2016
From: jmschreiber91 at gmail.com (Jacob Schreiber)
Date: Tue, 31 May 2016 11:18:23 -0700
Subject: [scikit-learn] Artificial neural network not learning lower
 values of the training sample
In-Reply-To: <CAGWVJJXtc2BTRuiPb=noPYXLhttNe0is+U4iZvaduiTCyZGu8g@mail.gmail.com>
References: <CAGWVJJWUJCFbyA7iKd8dK3qQe5nC2xu1Hm_ox_S-XYPKNkt5-Q@mail.gmail.com>
 <3FB2A8F0-9EAE-48CB-A0AC-98D9F27EA5A8@icloud.com>
 <CAGWVJJX47yEJF1g7DgSSwX5ChwBH6SXouDB5duo8yZMaBqeRfA@mail.gmail.com>
 <CA+ad8EuEb7fLAWgL+ne3D0n+41u_sV-NMFZtNtAzt-LqgRy73w@mail.gmail.com>
 <CAGWVJJXtc2BTRuiPb=noPYXLhttNe0is+U4iZvaduiTCyZGu8g@mail.gmail.com>
Message-ID: <CA+ad8EsNOPVDgGv_3To=vzv2WWcc=wj0gVQ616ivXXgUaxuJyA@mail.gmail.com>

Using the same feature set? How well do other estimators work? (Linear
regression, gradient boosting, etc...)

On Tue, May 31, 2016 at 11:10 AM, muhammad waseem <m.waseem.ahmad at gmail.com>
wrote:

> This problem has been solved in the literature before, I can post papers.
>
> On Tue, May 31, 2016 at 7:07 PM, Jacob Schreiber <jmschreiber91 at gmail.com>
> wrote:
>
>> Do you have any other baselines which you can compare to? It might be
>> helpful in seeing if this is a problem which can be learned.
>>
>> On Tue, May 31, 2016 at 10:47 AM, muhammad waseem <
>> m.waseem.ahmad at gmail.com> wrote:
>>
>>> Thanks for your reply. I have day, month, hour, temp, relative humidity,
>>> Wind speed as my input variables. I can't think of any other dependant
>>> variables. It is quite strange to me that I don't get results after using
>>> these input variables.
>>>
>>> On Tue, May 31, 2016 at 4:59 PM, Andrew Holmes <
>>> andrewholmes82 at icloud.com> wrote:
>>>
>>>> If the problem is that it?s confusing day and night, are you including
>>>> time of day as a parameter?
>>>>
>>>> Best wishes
>>>> Andrew
>>>>
>>>> @andrewholmes82 <http://twitter.com/andrewholmes82>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> On 31 May 2016, at 16:55, muhammad waseem <m.waseem.ahmad at gmail.com>
>>>> wrote:
>>>>
>>>> Hi All,
>>>> I am trying to train an ANN but until now it is not learning the lower
>>>> values of the training sample. I have tried using different python
>>>> libraries to train ANN. The aim is to predict solar radiation from other
>>>> weather parameters (regression problem). I think the ANN is confusing lower
>>>> values (winter/cloudy days) with the night-time values (probably). I have
>>>> tried the following but none of them worked;
>>>>
>>>>  1. Scaling data between different values e.g. [0,1],[-1,1]
>>>>  2. Standardising data to have zero mean and unit variance
>>>>  3. Shuffling the data
>>>>  4. Increasing the training samples (from 3 years to 10 years)
>>>>  5. Using different train functions
>>>>  6. Trying different transfer functions
>>>>  7. Using few input variables
>>>>  8. Varying hidden layers and hidden layers' neurons
>>>>
>>>> Any idea what could be wrong or any directions to try?
>>>>
>>>> Thanks
>>>> Kindest Regards
>>>> Waseem
>>>> _______________________________________________
>>>> scikit-learn mailing list
>>>> scikit-learn at python.org
>>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> scikit-learn mailing list
>>>> scikit-learn at python.org
>>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>>
>>>>
>>>
>>> _______________________________________________
>>> scikit-learn mailing list
>>> scikit-learn at python.org
>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>
>>>
>>
>> _______________________________________________
>> scikit-learn mailing list
>> scikit-learn at python.org
>> https://mail.python.org/mailman/listinfo/scikit-learn
>>
>>
>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160531/7bdd00ad/attachment-0001.html>

From m.waseem.ahmad at gmail.com  Tue May 31 14:52:26 2016
From: m.waseem.ahmad at gmail.com (muhammad waseem)
Date: Tue, 31 May 2016 19:52:26 +0100
Subject: [scikit-learn] Artificial neural network not learning lower
 values of the training sample
In-Reply-To: <CA+ad8EsNOPVDgGv_3To=vzv2WWcc=wj0gVQ616ivXXgUaxuJyA@mail.gmail.com>
References: <CAGWVJJWUJCFbyA7iKd8dK3qQe5nC2xu1Hm_ox_S-XYPKNkt5-Q@mail.gmail.com>
 <3FB2A8F0-9EAE-48CB-A0AC-98D9F27EA5A8@icloud.com>
 <CAGWVJJX47yEJF1g7DgSSwX5ChwBH6SXouDB5duo8yZMaBqeRfA@mail.gmail.com>
 <CA+ad8EuEb7fLAWgL+ne3D0n+41u_sV-NMFZtNtAzt-LqgRy73w@mail.gmail.com>
 <CAGWVJJXtc2BTRuiPb=noPYXLhttNe0is+U4iZvaduiTCyZGu8g@mail.gmail.com>
 <CA+ad8EsNOPVDgGv_3To=vzv2WWcc=wj0gVQ616ivXXgUaxuJyA@mail.gmail.com>
Message-ID: <CAGWVJJXOX_YZA-V=7aOfQQfGGJvL44sP6_D8mgSX1cf9jntq-w@mail.gmail.com>

I have tried Random forest (with gridseacrhCV) but did not get good
results.

On Tue, May 31, 2016 at 7:18 PM, Jacob Schreiber <jmschreiber91 at gmail.com>
wrote:

> Using the same feature set? How well do other estimators work? (Linear
> regression, gradient boosting, etc...)
>
> On Tue, May 31, 2016 at 11:10 AM, muhammad waseem <
> m.waseem.ahmad at gmail.com> wrote:
>
>> This problem has been solved in the literature before, I can post papers.
>>
>> On Tue, May 31, 2016 at 7:07 PM, Jacob Schreiber <jmschreiber91 at gmail.com
>> > wrote:
>>
>>> Do you have any other baselines which you can compare to? It might be
>>> helpful in seeing if this is a problem which can be learned.
>>>
>>> On Tue, May 31, 2016 at 10:47 AM, muhammad waseem <
>>> m.waseem.ahmad at gmail.com> wrote:
>>>
>>>> Thanks for your reply. I have day, month, hour, temp, relative
>>>> humidity, Wind speed as my input variables. I can't think of any other
>>>> dependant variables. It is quite strange to me that I don't get results
>>>> after using these input variables.
>>>>
>>>> On Tue, May 31, 2016 at 4:59 PM, Andrew Holmes <
>>>> andrewholmes82 at icloud.com> wrote:
>>>>
>>>>> If the problem is that it?s confusing day and night, are you including
>>>>> time of day as a parameter?
>>>>>
>>>>> Best wishes
>>>>> Andrew
>>>>>
>>>>> @andrewholmes82 <http://twitter.com/andrewholmes82>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> On 31 May 2016, at 16:55, muhammad waseem <m.waseem.ahmad at gmail.com>
>>>>> wrote:
>>>>>
>>>>> Hi All,
>>>>> I am trying to train an ANN but until now it is not learning the lower
>>>>> values of the training sample. I have tried using different python
>>>>> libraries to train ANN. The aim is to predict solar radiation from other
>>>>> weather parameters (regression problem). I think the ANN is confusing lower
>>>>> values (winter/cloudy days) with the night-time values (probably). I have
>>>>> tried the following but none of them worked;
>>>>>
>>>>>  1. Scaling data between different values e.g. [0,1],[-1,1]
>>>>>  2. Standardising data to have zero mean and unit variance
>>>>>  3. Shuffling the data
>>>>>  4. Increasing the training samples (from 3 years to 10 years)
>>>>>  5. Using different train functions
>>>>>  6. Trying different transfer functions
>>>>>  7. Using few input variables
>>>>>  8. Varying hidden layers and hidden layers' neurons
>>>>>
>>>>> Any idea what could be wrong or any directions to try?
>>>>>
>>>>> Thanks
>>>>> Kindest Regards
>>>>> Waseem
>>>>> _______________________________________________
>>>>> scikit-learn mailing list
>>>>> scikit-learn at python.org
>>>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>>>
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> scikit-learn mailing list
>>>>> scikit-learn at python.org
>>>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>>>
>>>>>
>>>>
>>>> _______________________________________________
>>>> scikit-learn mailing list
>>>> scikit-learn at python.org
>>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>>
>>>>
>>>
>>> _______________________________________________
>>> scikit-learn mailing list
>>> scikit-learn at python.org
>>> https://mail.python.org/mailman/listinfo/scikit-learn
>>>
>>>
>>
>> _______________________________________________
>> scikit-learn mailing list
>> scikit-learn at python.org
>> https://mail.python.org/mailman/listinfo/scikit-learn
>>
>>
>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160531/709e9ba4/attachment.html>

From andrewholmes82 at icloud.com  Tue May 31 13:54:45 2016
From: andrewholmes82 at icloud.com (Andrew Holmes)
Date: Tue, 31 May 2016 18:54:45 +0100
Subject: [scikit-learn] Artificial neural network not learning lower
 values of the training sample
In-Reply-To: <3FB2A8F0-9EAE-48CB-A0AC-98D9F27EA5A8@icloud.com>
References: <CAGWVJJWUJCFbyA7iKd8dK3qQe5nC2xu1Hm_ox_S-XYPKNkt5-Q@mail.gmail.com>
 <3FB2A8F0-9EAE-48CB-A0AC-98D9F27EA5A8@icloud.com>
Message-ID: <4A030BCE-4A5F-42F6-B7DB-40D0D388EBBD@icloud.com>

When you say it?s not learning ?lower values?, does that mean the model has good predictions on high values in the test set, but poor performance on the low ones?

Have you tried simpler models like tree, random forest and svm as a benchmark?

Best wishes
Andrew

@andrewholmes82 <http://twitter.com/andrewholmes82>








> On 31 May 2016, at 16:59, Andrew Holmes <andrewholmes82 at icloud.com> wrote:
> 
> If the problem is that it?s confusing day and night, are you including time of day as a parameter?
> 
> Best wishes
> Andrew
> 
> @andrewholmes82 <http://twitter.com/andrewholmes82>
> 
> 
> 
> 
> 
> 
> 
> 
>> On 31 May 2016, at 16:55, muhammad waseem <m.waseem.ahmad at gmail.com <mailto:m.waseem.ahmad at gmail.com>> wrote:
>> 
>> Hi All,
>> I am trying to train an ANN but until now it is not learning the lower values of the training sample. I have tried using different python libraries to train ANN. The aim is to predict solar radiation from other weather parameters (regression problem). I think the ANN is confusing lower values (winter/cloudy days) with the night-time values (probably). I have tried the following but none of them worked;
>> 
>>  1. Scaling data between different values e.g. [0,1],[-1,1]
>>  2. Standardising data to have zero mean and unit variance
>>  3. Shuffling the data
>>  4. Increasing the training samples (from 3 years to 10 years)
>>  5. Using different train functions
>>  6. Trying different transfer functions
>>  7. Using few input variables
>>  8. Varying hidden layers and hidden layers' neurons
>> 
>> Any idea what could be wrong or any directions to try?
>> 
>> Thanks
>> Kindest Regards
>> Waseem
>> _______________________________________________
>> scikit-learn mailing list
>> scikit-learn at python.org <mailto:scikit-learn at python.org>
>> https://mail.python.org/mailman/listinfo/scikit-learn
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160531/2009ad3b/attachment-0001.html>

From m.waseem.ahmad at gmail.com  Tue May 31 15:00:08 2016
From: m.waseem.ahmad at gmail.com (muhammad waseem)
Date: Tue, 31 May 2016 20:00:08 +0100
Subject: [scikit-learn] Artificial neural network not learning lower
 values of the training sample
In-Reply-To: <4A030BCE-4A5F-42F6-B7DB-40D0D388EBBD@icloud.com>
References: <CAGWVJJWUJCFbyA7iKd8dK3qQe5nC2xu1Hm_ox_S-XYPKNkt5-Q@mail.gmail.com>
 <3FB2A8F0-9EAE-48CB-A0AC-98D9F27EA5A8@icloud.com>
 <4A030BCE-4A5F-42F6-B7DB-40D0D388EBBD@icloud.com>
Message-ID: <CAGWVJJVDp_rFwK=JjyR8_34_RkFO_DUWGno0AHTsaR28BbdHMA@mail.gmail.com>

Yes, it has poor performance (higher errors) on lower values.
I have tried random forest but as I mentioned it did not give good results
either, I can try SVR.

Kindest Regards
Waseem

On Tue, May 31, 2016 at 6:54 PM, Andrew Holmes <andrewholmes82 at icloud.com>
wrote:

> When you say it?s not learning ?lower values?, does that mean the model
> has good predictions on high values in the test set, but poor performance
> on the low ones?
>
> Have you tried simpler models like tree, random forest and svm as a
> benchmark?
>
> Best wishes
> Andrew
>
> @andrewholmes82 <http://twitter.com/andrewholmes82>
>
>
>
>
>
>
>
>
> On 31 May 2016, at 16:59, Andrew Holmes <andrewholmes82 at icloud.com> wrote:
>
> If the problem is that it?s confusing day and night, are you including
> time of day as a parameter?
>
> Best wishes
> Andrew
>
> @andrewholmes82 <http://twitter.com/andrewholmes82>
>
>
>
>
>
>
>
>
> On 31 May 2016, at 16:55, muhammad waseem <m.waseem.ahmad at gmail.com>
> wrote:
>
> Hi All,
> I am trying to train an ANN but until now it is not learning the lower
> values of the training sample. I have tried using different python
> libraries to train ANN. The aim is to predict solar radiation from other
> weather parameters (regression problem). I think the ANN is confusing lower
> values (winter/cloudy days) with the night-time values (probably). I have
> tried the following but none of them worked;
>
>  1. Scaling data between different values e.g. [0,1],[-1,1]
>  2. Standardising data to have zero mean and unit variance
>  3. Shuffling the data
>  4. Increasing the training samples (from 3 years to 10 years)
>  5. Using different train functions
>  6. Trying different transfer functions
>  7. Using few input variables
>  8. Varying hidden layers and hidden layers' neurons
>
> Any idea what could be wrong or any directions to try?
>
> Thanks
> Kindest Regards
> Waseem
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>
>
>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160531/58349d30/attachment.html>

From andrewholmes82 at icloud.com  Tue May 31 15:01:45 2016
From: andrewholmes82 at icloud.com (Andrew Holmes)
Date: Tue, 31 May 2016 20:01:45 +0100
Subject: [scikit-learn] Artificial neural network not learning lower
 values of the training sample
In-Reply-To: <CAGWVJJVDp_rFwK=JjyR8_34_RkFO_DUWGno0AHTsaR28BbdHMA@mail.gmail.com>
References: <CAGWVJJWUJCFbyA7iKd8dK3qQe5nC2xu1Hm_ox_S-XYPKNkt5-Q@mail.gmail.com>
 <3FB2A8F0-9EAE-48CB-A0AC-98D9F27EA5A8@icloud.com>
 <4A030BCE-4A5F-42F6-B7DB-40D0D388EBBD@icloud.com>
 <CAGWVJJVDp_rFwK=JjyR8_34_RkFO_DUWGno0AHTsaR28BbdHMA@mail.gmail.com>
Message-ID: <DEA9FCCD-26BA-462A-B028-392D1FF9A49B@icloud.com>

Is the training set unbalanced between high and low values? Ie, many more of the high ones?

Best wishes
Andrew

@andrewholmes82 <http://twitter.com/andrewholmes82>








> On 31 May 2016, at 20:00, muhammad waseem <m.waseem.ahmad at gmail.com> wrote:
> 
> Yes, it has poor performance (higher errors) on lower values. 
> I have tried random forest but as I mentioned it did not give good results either, I can try SVR.
> 
> Kindest Regards
> Waseem
> 
> On Tue, May 31, 2016 at 6:54 PM, Andrew Holmes <andrewholmes82 at icloud.com <mailto:andrewholmes82 at icloud.com>> wrote:
> When you say it?s not learning ?lower values?, does that mean the model has good predictions on high values in the test set, but poor performance on the low ones?
> 
> Have you tried simpler models like tree, random forest and svm as a benchmark?
> 
> Best wishes
> Andrew
> 
> @andrewholmes82 <http://twitter.com/andrewholmes82>
> 
> 
> 
> 
> 
> 
> 
> 
>> On 31 May 2016, at 16:59, Andrew Holmes <andrewholmes82 at icloud.com <mailto:andrewholmes82 at icloud.com>> wrote:
>> 
>> If the problem is that it?s confusing day and night, are you including time of day as a parameter?
>> 
>> Best wishes
>> Andrew
>> 
>> @andrewholmes82 <http://twitter.com/andrewholmes82>
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>>> On 31 May 2016, at 16:55, muhammad waseem <m.waseem.ahmad at gmail.com <mailto:m.waseem.ahmad at gmail.com>> wrote:
>>> 
>>> Hi All,
>>> I am trying to train an ANN but until now it is not learning the lower values of the training sample. I have tried using different python libraries to train ANN. The aim is to predict solar radiation from other weather parameters (regression problem). I think the ANN is confusing lower values (winter/cloudy days) with the night-time values (probably). I have tried the following but none of them worked;
>>> 
>>>  1. Scaling data between different values e.g. [0,1],[-1,1]
>>>  2. Standardising data to have zero mean and unit variance
>>>  3. Shuffling the data
>>>  4. Increasing the training samples (from 3 years to 10 years)
>>>  5. Using different train functions
>>>  6. Trying different transfer functions
>>>  7. Using few input variables
>>>  8. Varying hidden layers and hidden layers' neurons
>>> 
>>> Any idea what could be wrong or any directions to try?
>>> 
>>> Thanks
>>> Kindest Regards
>>> Waseem
>>> _______________________________________________
>>> scikit-learn mailing list
>>> scikit-learn at python.org <mailto:scikit-learn at python.org>
>>> https://mail.python.org/mailman/listinfo/scikit-learn <https://mail.python.org/mailman/listinfo/scikit-learn>
>> 
> 
> 
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org <mailto:scikit-learn at python.org>
> https://mail.python.org/mailman/listinfo/scikit-learn <https://mail.python.org/mailman/listinfo/scikit-learn>
> 
> 
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160531/fa29be98/attachment-0001.html>

From m.waseem.ahmad at gmail.com  Tue May 31 15:02:50 2016
From: m.waseem.ahmad at gmail.com (muhammad waseem)
Date: Tue, 31 May 2016 20:02:50 +0100
Subject: [scikit-learn] Artificial neural network not learning lower
 values of the training sample
In-Reply-To: <CAGWVJJVDp_rFwK=JjyR8_34_RkFO_DUWGno0AHTsaR28BbdHMA@mail.gmail.com>
References: <CAGWVJJWUJCFbyA7iKd8dK3qQe5nC2xu1Hm_ox_S-XYPKNkt5-Q@mail.gmail.com>
 <3FB2A8F0-9EAE-48CB-A0AC-98D9F27EA5A8@icloud.com>
 <4A030BCE-4A5F-42F6-B7DB-40D0D388EBBD@icloud.com>
 <CAGWVJJVDp_rFwK=JjyR8_34_RkFO_DUWGno0AHTsaR28BbdHMA@mail.gmail.com>
Message-ID: <CAGWVJJXwNB9LUc-EQ=+KG4RiQeH-aY5KR5tXFXyO_9ZMBSSrLA@mail.gmail.com>

Please find below results for ANN
. [image: Inline image 1]

On Tue, May 31, 2016 at 8:00 PM, muhammad waseem <m.waseem.ahmad at gmail.com>
wrote:

> Yes, it has poor performance (higher errors) on lower values.
> I have tried random forest but as I mentioned it did not give good results
> either, I can try SVR.
>
> Kindest Regards
> Waseem
>
> On Tue, May 31, 2016 at 6:54 PM, Andrew Holmes <andrewholmes82 at icloud.com>
> wrote:
>
>> When you say it?s not learning ?lower values?, does that mean the model
>> has good predictions on high values in the test set, but poor performance
>> on the low ones?
>>
>> Have you tried simpler models like tree, random forest and svm as a
>> benchmark?
>>
>> Best wishes
>> Andrew
>>
>> @andrewholmes82 <http://twitter.com/andrewholmes82>
>>
>>
>>
>>
>>
>>
>>
>>
>> On 31 May 2016, at 16:59, Andrew Holmes <andrewholmes82 at icloud.com>
>> wrote:
>>
>> If the problem is that it?s confusing day and night, are you including
>> time of day as a parameter?
>>
>> Best wishes
>> Andrew
>>
>> @andrewholmes82 <http://twitter.com/andrewholmes82>
>>
>>
>>
>>
>>
>>
>>
>>
>> On 31 May 2016, at 16:55, muhammad waseem <m.waseem.ahmad at gmail.com>
>> wrote:
>>
>> Hi All,
>> I am trying to train an ANN but until now it is not learning the lower
>> values of the training sample. I have tried using different python
>> libraries to train ANN. The aim is to predict solar radiation from other
>> weather parameters (regression problem). I think the ANN is confusing lower
>> values (winter/cloudy days) with the night-time values (probably). I have
>> tried the following but none of them worked;
>>
>>  1. Scaling data between different values e.g. [0,1],[-1,1]
>>  2. Standardising data to have zero mean and unit variance
>>  3. Shuffling the data
>>  4. Increasing the training samples (from 3 years to 10 years)
>>  5. Using different train functions
>>  6. Trying different transfer functions
>>  7. Using few input variables
>>  8. Varying hidden layers and hidden layers' neurons
>>
>> Any idea what could be wrong or any directions to try?
>>
>> Thanks
>> Kindest Regards
>> Waseem
>> _______________________________________________
>> scikit-learn mailing list
>> scikit-learn at python.org
>> https://mail.python.org/mailman/listinfo/scikit-learn
>>
>>
>>
>>
>> _______________________________________________
>> scikit-learn mailing list
>> scikit-learn at python.org
>> https://mail.python.org/mailman/listinfo/scikit-learn
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160531/2e5d2881/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Results.png
Type: image/png
Size: 164056 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160531/2e5d2881/attachment-0001.png>

From m.waseem.ahmad at gmail.com  Tue May 31 15:05:07 2016
From: m.waseem.ahmad at gmail.com (muhammad waseem)
Date: Tue, 31 May 2016 20:05:07 +0100
Subject: [scikit-learn] Artificial neural network not learning lower
 values of the training sample
In-Reply-To: <DEA9FCCD-26BA-462A-B028-392D1FF9A49B@icloud.com>
References: <CAGWVJJWUJCFbyA7iKd8dK3qQe5nC2xu1Hm_ox_S-XYPKNkt5-Q@mail.gmail.com>
 <3FB2A8F0-9EAE-48CB-A0AC-98D9F27EA5A8@icloud.com>
 <4A030BCE-4A5F-42F6-B7DB-40D0D388EBBD@icloud.com>
 <CAGWVJJVDp_rFwK=JjyR8_34_RkFO_DUWGno0AHTsaR28BbdHMA@mail.gmail.com>
 <DEA9FCCD-26BA-462A-B028-392D1FF9A49B@icloud.com>
Message-ID: <CAGWVJJVBuA6Q29eQguw8iq3c0337j-eYiLJpG0JyVBRGQNW2zw@mail.gmail.com>

I try to balance it out, the dataset is very periodic type (similar
behaviour in an year)

On Tue, May 31, 2016 at 8:01 PM, Andrew Holmes <andrewholmes82 at icloud.com>
wrote:

> Is the training set unbalanced between high and low values? Ie, many more
> of the high ones?
>
> Best wishes
> Andrew
>
> @andrewholmes82 <http://twitter.com/andrewholmes82>
>
>
>
>
>
>
>
>
> On 31 May 2016, at 20:00, muhammad waseem <m.waseem.ahmad at gmail.com>
> wrote:
>
> Yes, it has poor performance (higher errors) on lower values.
> I have tried random forest but as I mentioned it did not give good results
> either, I can try SVR.
>
> Kindest Regards
> Waseem
>
> On Tue, May 31, 2016 at 6:54 PM, Andrew Holmes <andrewholmes82 at icloud.com>
> wrote:
>
>> When you say it?s not learning ?lower values?, does that mean the model
>> has good predictions on high values in the test set, but poor performance
>> on the low ones?
>>
>> Have you tried simpler models like tree, random forest and svm as a
>> benchmark?
>>
>> Best wishes
>> Andrew
>>
>> @andrewholmes82 <http://twitter.com/andrewholmes82>
>>
>>
>>
>>
>>
>>
>>
>>
>> On 31 May 2016, at 16:59, Andrew Holmes <andrewholmes82 at icloud.com>
>> wrote:
>>
>> If the problem is that it?s confusing day and night, are you including
>> time of day as a parameter?
>>
>> Best wishes
>> Andrew
>>
>> @andrewholmes82 <http://twitter.com/andrewholmes82>
>>
>>
>>
>>
>>
>>
>>
>>
>> On 31 May 2016, at 16:55, muhammad waseem <m.waseem.ahmad at gmail.com>
>> wrote:
>>
>> Hi All,
>> I am trying to train an ANN but until now it is not learning the lower
>> values of the training sample. I have tried using different python
>> libraries to train ANN. The aim is to predict solar radiation from other
>> weather parameters (regression problem). I think the ANN is confusing lower
>> values (winter/cloudy days) with the night-time values (probably). I have
>> tried the following but none of them worked;
>>
>>  1. Scaling data between different values e.g. [0,1],[-1,1]
>>  2. Standardising data to have zero mean and unit variance
>>  3. Shuffling the data
>>  4. Increasing the training samples (from 3 years to 10 years)
>>  5. Using different train functions
>>  6. Trying different transfer functions
>>  7. Using few input variables
>>  8. Varying hidden layers and hidden layers' neurons
>>
>> Any idea what could be wrong or any directions to try?
>>
>> Thanks
>> Kindest Regards
>> Waseem
>> _______________________________________________
>> scikit-learn mailing list
>> scikit-learn at python.org
>> https://mail.python.org/mailman/listinfo/scikit-learn
>>
>>
>>
>>
>> _______________________________________________
>> scikit-learn mailing list
>> scikit-learn at python.org
>> https://mail.python.org/mailman/listinfo/scikit-learn
>>
>>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>
>
> _______________________________________________
> scikit-learn mailing list
> scikit-learn at python.org
> https://mail.python.org/mailman/listinfo/scikit-learn
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20160531/b864e41e/attachment.html>