Fig. 7From: Exploring kinase family inhibitors and their moiety preferences using deep SHapley additive exPlanationsThe DNN model framework for predicting kinase family inhibitors. The input layer includes 238 nodes which are the moiety features of an inhibitor. The four hidden layers comprise 1024, 768, 512, and 256 neurons (nodes), respectively. The output layer uses sigmoid to turn the model’s prediction into [0, 1], to make it a probabilistic estimateBack to article page