I'm trying to do some supervised learning using scikit-learn MLPRegressor. I want to know what is the leverage of numbers of epoch in supervised learning. But when I set the MLP like this : mlp = MLPRegressor(max_iter=100, learning_rate_init=0.1) and then this : mlp = MLPRegressor(max_iter=200, learning_rate_init=0.1) and then this : mlp = MLPRegressor(max_iter=500, learning_rate_init=0.1) The prediction scores remain the same. I don't know if it's correct or not for using max_iter to set the numbers of epoch, because prediction scores are constant although I change the numbers of max_iter. But when I change the learning rate, the scores are change so there are some leverages of learning rate. Anyone can help? thanks Continue reading...
