Table 9 Proposed model performance comparison—with and without PCA
From: Short-term stock market price trend prediction using a comprehensive deep learning system
Metrics name | LSTM trained on 29 features | LSTM trained on 5 principal components |
|---|---|---|
Loss | 0.0702 | 0.0848 |
F1 score | 0.9323 | 0.9194 |
Binary accuracy | 0.9325 | 0.9193 |
MSE | 0.0669 | 0.0772 |
MAE | 0.0702 | 0.0848 |
TPR | 0.96 | 0.92 |
TNR | 0.90 | 0.91 |
FPR | 0.04 | 0.08 |
FNR | 0.10 | 0.09 |