Skip to main content

Table 7 Table summarizes the accuracy of ANN, GRU, BILSTM, LSTM, and 1DCNN model for 32 SNPs

From: Eye-color and Type-2 diabetes phenotype prediction from genotype data using deep learning methods

Model, SNPs = 32 Activation Dropout Optimizer Batchsize Epochs Validation Accuracy
ANN Softmax 0.3 RMSprop 10 100 0.2 0.91
Softmax 0.3 SGD 1 100 0.2 0.91
Relu 0.3 RMSprop 15 50 0.2 0.91
Relu 0.3 SGD 20 100 0.2 0.92
Relu 0.3 SGD 1 100 0.3 0.92
GRU Sigmoid 0.2 Adam 1 10 0.2 0.92
Sigmoid 0.3 Adam 1 50 0.2 0.91
Sigmoid 0.2 SGD 1 20 0.2 0.91
BILSTM Sigmoid 0.2 RMSprop 15 20 0.2 0.91
1DCNN Sigmoid 0.2 Adam 1 20 0.3 0.92
Sigmoid 0.2 Adam 15 100 0.3 0.92
Softmax 0.3 Adam 10 50 0.2 0.91
Softmax 0.2 RMSprop 20 100 0.2 0.91
Relu 0.3 RMSprop 10 50 0.2 0.91
  1. ANN, GRU, and 1DCNN perform well with an accuracy of 0.92%