Fig. 2From: A biological network-based regularized artificial neural network model for robust phenotype prediction from gene expression dataFigure represents a gene regulatory network based ANN. The input layer corresponds to genes, while the hidden layer correspond to regulators. The connections between the input and the hidden layer are based on regulatory interactions. The ridge â„“ 2 regularization is applied on these connection. The output layer consists of a single node for binary classification. The nodes in hidden layer are fully connected to the output node. The â„“ 1 regularization is applied to these connectionsBack to article page