Change default Stage of PSL when predicting (currently last stage is used regardless of performance)

[JonasHanselle]

The default model which is used when "psl.predict()" is called shall be the best performing one in terms of the stage loss (expected entropy per default). Currently, it is the model that is the last one in the list, which may not be the best performing one.