Hi scikit,I coded Gaussian RBM in scikit form. Testing the code, however, on the digits dataset gave me unreliable performance. Though I followed references and py2learn's G-RBM to the T, the scores suggest that my implementation is wrong. Applying logistic regression on the raw pixels gave score of 0.947, whereas on G-RBM features the score is 0.419. Seemingly, the code is incorrect or G-RBM is not suitable for scikit's digits dataset.
I described the issue on http://metaoptimize.com <http://metaoptimize.com/> [1] more extensively, but didn't get a comment yet. So, I figured I could nudge scikit for help ;).
If you wish to follow up on the problem please see the link below. It has the code I developed to aid your investigation.
[1] http://metaoptimize.com/qa/questions/14068/explanation-for-why-this-gaussian-rbm-did-not-perform-so-well
------------------------------------------------------------------------------ October Webinars: Code for Performance Free Intel webinars can help you accelerate application performance. Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from the latest Intel processors and coprocessors. See abstracts and register > http://pubads.g.doubleclick.net/gampad/clk?id=60133471&iu=/4140/ostg.clktrk
_______________________________________________ Scikit-learn-general mailing list Scikit-learn-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
