2011/12/28 Nick Wilson <[email protected]>: > On Tue, Dec 27, 2011 at 6:23 PM, Olivier Grisel > <[email protected]> wrote: >> Hi all, >> >> I came across the following blog post about Andrew Ng's ML class and I >> like the training / validation errors plots to find out whether the >> model is too biased (underfitting) or two lax (high variance, >> over-fitting). >> >> http://digitheadslabnotebook.blogspot.com/2011/12/practical-advice-for-applying-machine.html > > I just completed this class. I took detailed notes (many screenshots) > for myself on the lessons that offered practical advice. They may be > useful to you: > > https://sites.google.com/a/njwilson.net/ml-class/practical > > Perhaps I should clean them up and post them for a wider audience... > >> Has someone tried to apply this trick in practice? Do you think it >> would be interesting to provide new utilities to easily draw such >> plots as a way to qualitatively check / validate model selection >> achieved by automated methods such as grid search CV? > > I'm quite new to machine learning and haven't had the chance to apply > it yet, so I can't comment on these techniques in practice. But as > someone who has completed the class and is playing with scikit-learn, > I would love to see support for easily generating these plots.
Thanks for putting your lecture notes online. BTW, here is a gist to play with this kind of plots on text classification data: https://gist.github.com/1540431 It seems to be an interesting tool to me. We need to find a non-trivial overfitting example that would run in an acceptable time with the datasets available in the scikit. -- Olivier http://twitter.com/ogrisel - http://github.com/ogrisel ------------------------------------------------------------------------------ Ridiculously easy VDI. With Citrix VDI-in-a-Box, you don't need a complex infrastructure or vast IT resources to deliver seamless, secure access to virtual desktops. With this all-in-one solution, easily deploy virtual desktops for less than the cost of PCs and save 60% on VDI infrastructure costs. Try it free! http://p.sf.net/sfu/Citrix-VDIinabox _______________________________________________ Scikit-learn-general mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
