Table 2.

Classifier Hyper-parameters.

Classifier	Parameter
DBN	learning rate:0.01 Max Iter = 50.0; Batch Size = 200; Step Ratio = 0.01; Drop Out Rate = 0.1; Verbose = ’true’; hidden neuron: 50 Initial Momentum = 0.5; % momentum for first five iterations Final Momentum = 0.9; % momentum for remaining iterations Weight Cost = 0.0002; % costs of weight update Initial Momentum Iter = 5; Max Iter = 100; Step Ratio = 0.01; BatchSize = 0; Verbose = false;
RNN	input layer-1 lstm layer-1 fullyConnectedLayer-3 activation: tanh ‘optimizer’: ‘sgdm’,… ‘MaxEpochs’,50, … ‘MiniBatchSize’,70, … ‘Verbose’, false