Closed gpestre closed 4 years ago
Rejected in favour of PR #90.
Since we don't have functionality where passing separate x_train
and y_train
is natural, getting rid of fit()
altogether and instead doing it in the constructor seems cleaner than expanding the arguably already superfluous fit()
method.
Update API for
DecisionTree
andRandomForest
to take test data in.fit
method instead of constructor.