DNN was established using Python and open-source library Keras from Tensorflow. Our DNN consist of three hidden layers and one output layers. The activation function used in the hidden layers is “relu” function and the activation used in the output layer is “softmax” function. The optimizer used for the DNN is “adam” optimizer for its better performance on big data. The data was divided into the batch size of ‘10’ and number of epochs used- ‘100’. Number of hidden layers used are ‘4’ followed by ‘binary cross-entropy’ loss function.
The datasets used in this study are available at - https://data.matr.io/1