Fig. 3From: Deep learning architectures for multi-label classification of intelligent health risk predictionA comparison of additional layers added to the MLP. The hyperparameters are: 1000 epochs, 0.1 learning rate, 35 hidden layer units, hidden layers from 1 to 10, and no dropout to one dropout layer to all dropout layers. These parameters were chosen because they gave the best overall performance for MLP with 1 or 2 layersBack to article page