First of all thanks Asa for this promising library.
It contains a lot of the features missing in libsvm (and most other svm packages) as e.g. stratifiedCV.
i am desperately trying to reduce the verbosity
(using paramGrid) but although I tried to set verbose=False e.g. in the cv method it seems to have little or no effect. Probably I am doing something stupid but some methods as e.g. SparseCDataSet.copy, WrapperDataSet.attachLabels etc. seem to print uncontrolled to stdout.
btw. the example for paramGrid seems to be a cut/paste error. For me it was not as simple as stated:
First to set gamma you need to assign a Gaussian classifier to the method. As assess is not in init.py one needs to import it and create a classifier with Gaussian kernel, which then can be passed to paramGrid.
thanks for your comments.
regarding verbosity -- some of it comes from the libsvm code itself, and some are my own debuggin code that i haven't gotten ridden of yet. you're right the verbosity flag of CV has very little effect.
you're also right about the grid example -- i modified it.
Log in to post a comment.
Sign up for the SourceForge newsletter:
You seem to have CSS turned off.
Please don't fill out this field.