Problem: MLPClassifier:ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached

MLPClassifier:ConvergenceWarning: Stochastic Optimizer: Maximum iterations (200) reached

Background: Accourding to the description of this problem, it means that this model can not be the best state for trainning data.

Solution: When you see warning, as the problem said, the iteration could be less. So, we could set the “max_iter=1000” , and restart the programme. Then, the warning disapperas rapidly.

Summary: When we learn more about the nerual network, we could not only focus on the overfitting but also the under-fitting. Although the best way to solve under-fitting is to increase the datasets, we should find the reason of under-fitting from datasets, itself.