I am using a MLP model for classification.
When I predict for new data, I want to keep only those predictions whose probability of prediction is larger than 0.5, and cha