[Baypiggies] Suggested Reading: Machine learning and Python

[Paul Ivanov]

Venkatraman S, on 2011-01-06 10:16,  wrote:
> On Thu, Jan 6, 2011 at 9:43 AM, Venkatraman S <venkat83 at gmail.com> wrote:
> 
> >
> > I have gone through Collective Intelligence and Python text Processing with
> > NLTK, and was wondering whether there is any suggested reading to get into
> > deeper depths in statistical machine learning - something which gives a
> > 'basic' introduction to mixture models , EM etc.
> >
> 
> Let me expatiate a little more before you point me to the Measuring Measure
> link or Andy Ng classes : I have seen them and have attended the first 3
> lectures of Andy. I also went through courses taught at MIT(via OCW) and
> Jordan's classes(UCB) - but most of these stuff are heavy theoretical - not
> that I am against theory, but i want some hands-on to understand how theory
> is implemented.
> For eg. PythonTextProcessingUsingNLTK does a great job in understanding
> various aspects of text processing by playing with text parallely. Is there
> something similar to understand kernel methods or mixture models?
> To give you one more idea, i asked this question in
> stackoverflow<http://stats.stackexchange.com/questions/5960/how-to-identify-a-bimodal-distribution>.Working(handson)
> on data while understanding/learning them is a great way to learn :)

I was the TA for Vision Science 265 - Bruno Olshausen's Neural
Computation course[1] this past semester at UCB: 

    This course provides an introduction to the theory of neural
    computation. The goal is to familiarize students with the
    major theoretical frameworks and models used in neuroscience
    and psychology, and to provide hands-on experience in using
    these models. Topics include neural network models,
    supervised and unsupervised learning rules, associative
    memory models, probabilistic/graphical models, sensorimotor
    loops, and models of neural coding in the brain.

It was the first year we allowed the students to do the "lab"
assignments in Python, and I wrote up the templates for most of
them. The assignments involve little toy data sets and boiler
plate code to get you going - most students find them pretty
engaging (I know I did when I took the course 4 years ago).
Videos for all of the lectures are up on Archive.org and linnked
from [1]. Check the syllabus for the topics covered.

Also, although I have not read it - there's Stephen Marsland's
_Machine Learning: An Algorithmic Perspective_ [2], which comes
with a lot of python code, as well.

1. http://redwood.berkeley.edu/wiki/Vs265
2. http://www-ist.massey.ac.nz/smarsland/MLbook.html

best,
-- 
Paul Ivanov
314 address only used for lists,  off-list direct email at:
http://pirsquared.org | GPG/PGP key id: 0x0F3E28F7 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://mail.python.org/pipermail/baypiggies/attachments/20110105/66b7f6b2/attachment-0001.pgp>