Table 5 Predictive performance metrics for each model.
Model | R2 | MSE | RMSE | MAE | MAPE | Training time |
|---|---|---|---|---|---|---|
CNN | 56.434% | 7.450 × 10−8 | 2.729 × 10−4 | 2.501 × 10−4 | 0.131% | 8.775 |
GRU | 58.177% | 5.645 × 10−8 | 2.376 × 10−4 | 1.769 × 10−4 | 0.093% | 10.886 |
CNN-BILSTM | 85.595% | 1.740 × 10−8 | 1.319 × 10−4 | 1.132 × 10−4 | 0.059% | 22.337 |
CNN-LSTM | 87.891% | 1.315 × 10−8 | 1.147 × 10−4 | 9.569 × 10−5 | 0.050% | 20.359 |
LSTM | 90.862% | 1.065 × 10−8 | 1.032 × 10−4 | 9.119 × 10−5 | 0.048% | 9.035 |
BILSTM | 91.786% | 9.157 × 10−9 | 9.569 × 10−5 | 6.999 × 10−5 | 0.037% | 10.177 |
BILSTM-ADABOOST | 95.483% | 5.038 × 10−9 | 7.098 × 10−5 | 5.917 × 10−5 | 0.031% | 31.094 |
