Data Sprint 50

using classifiers + optuna "as usual". I don't think optuna made sense. I wonder that scores are scattered in spite of such a dataset which needs less preprocesses.