[FIXED] Validation_data argument in sklearn fit method

Issue

I am trying to get a separate predefined validation set passed to sklearn fit method. I was wondering if there is already a given way in the sklearn API to fulfill this need, like it is given within the keras API:

model.fit(train_X, train_y, validation_data=(val_x, val_y))

and

model.fit(train_X, train_y, validation_split=0.3)

Is there an equivialent within the sklearn API for the argument validation_data=(x_val, y_val)? I only can find the validation_split argument within the documentation.

Solution

according to the docs here, no, you cannot, but you can use the score method after training on the validation set - in this case val_x and val_y - to see the performance of the model

Answered By - Mahmoud Youssef