Hello, I have a custom searchlight function that uses SVM as a classifier. The training and testing data have four features that were differently attended in a task. I want to train a classifier to recognize all four features, and then test it on the testing data. As a result of the testing, I want the classifier to give me an array of probabilities for how good each feature was tested. The array then should have something like four columns, one for each feature, and a row for each volume (test trial).
I tried to use lsvm.ca<http://lsvm.ca>_enable[‘estimate’] and lsvm.ca<http://lsvm.ca>.estimates to get the probabilities, but it gives me something else, but probabilities (btw, lsvm is my classifier). Please let me know if anything does not make sense. I just started with mvpa and can explain my goals/methods somewhat obscure. Best wishes, Dmitrii ------------------------------------------------ Dmitrii Paniukov Doctoral Student and Research Assistant Experimental/Applied Cognitive Psychology Texas Tech University Email: [email protected]<mailto:[email protected]> ------------------------------------------------
_______________________________________________ Pkg-ExpPsy-PyMVPA mailing list [email protected] http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-exppsy-pymvpa
