[scikit-learn] Any clustering algo to cluster multiple timing series data?

[Mikkel Haggren Brynildsen]

You can use it to get a single similarity / closeness number between two timeseries and then feed that into a clustering algorithm.

For instance look at
https://github.com/markdregan/K-Nearest-Neighbors-with-Dynamic-Time-Warping


as a first idea:
if you expand the distance function d = lambda x,y: abs(x-y) to a multivariate local distance

d2 = lambda a,b: np.sqrt(float((a[0]-b[0])**2 + (a[1]-b[1])**2)
(or any other n-dim metric)

Then you have an algorithm that could cluster the timeseries.

It does also work when the timeseries are of equal length…

Best
Mikkel Brynildsen


From: scikit-learn <scikit-learn-bounces+mbrynildsen=grundfos.com at python.org> On Behalf Of lampahome
Sent: 17. januar 2019 08:45
To: Scikit-learn mailing list <scikit-learn at python.org>
Subject: Re: [scikit-learn] Any clustering algo to cluster multiple timing series data?



Mikkel Haggren Brynildsen <mbrynildsen at grundfos.com<mailto:mbrynildsen at grundfos.com>> 於 2019年1月17日 週四 下午3:07寫道：
What about dynamic time warping ?

I thought DTW is used to different length of two datasets
But I only get the same length of two datasets.
Maybe it doesn't work?


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scikit-learn/attachments/20190117/09304db8/attachment-0001.html>