Bug with custom classifier

[xiaolongwu0713]

If you uncomment below, and run your test_file_multi.py, you will know:
FBCSP_multi_clf = FBCSP_Multiclass(trials_dict, fs, classifier = SVC(kernel = 'rbf', probability = True), print_var = print_var)

[TekinGunasar]

I have an implementation that fixes this issue. It occurs because every time he trained a new classifier for the One vs Rest system, it would overwrite the previous one, which is an issue in itself, but especially an issue here when the MIBIF feature extraction method results in a variable number of features per classifier. It’s a simple fix - you just have to make a copy of the SkLearn classifier object.