Here I tested a different way of doing hyperparameter optimization in one-vs-rest classifiers.

The idea is that sklearn.multiclass.OneVsRestClassifier uses the same hyperparameter value for all base estimators. I was curious what happens if you allowed every estimator to have it's own hyperparameter value.

It's quite likely that this is theoretically unsound, but I was curious.

I tested it on internal Washington Post data and on the 20 Newsgroups data set. At best it seems to perform just as well as the same hyperparameter for all estimators.