neurospin / pylearn-parsimony

Structured Machine Learning in Python
BSD 3-Clause "New" or "Revised" License
45 stars 16 forks source link

INFO_PROVIDED for the static Conesta algorithm #6

Closed FH235918 closed 9 years ago

FH235918 commented 9 years ago

When tying to add a unit test for the static Conesta, I noticed that using the gap is a very good idea to test whether the algorithm converged correctly. This is especially useful when testing the logistic regression model since in this case no explicit solution is known. So, I added Info.gap to the satic_Conesta 's INFO_PROVIDED in the proximal file.

duchesnay commented 9 years ago

done