Research Article
Water Quality Prediction Based on Hybrid Deep Learning Algorithm
Table 1
Performance metrics of the variable methods.
| Features | Method | RMSE | MAE | MAPE | R2 |
| Dissolved oxygen | ANN | 0.247 | 0.199 | 0.175 | 0.91 | BPNN | 0.179 | 0.139 | 0.119 | 0.90 | RNN | 0.135 | 0.092 | 0.079 | 0.89 | LSTM–GWO–FSO | 0.083 | 0.055 | 0.044 | 0.94 |
| COD | ANN | 0.059 | 0.038 | 0.291 | 0.902 | BPNN | 0.049 | 0.029 | 0.191 | 0.91 | RNN | 0.039 | 0.021 | 0.142 | 0.93 | LSTM–GWO–FSO | 0.016 | 0.011 | 0.080 | 0.95 |
| NH3─N | ANN | 0.015 | 0.014 | 0.598 | 0.92 | BPNN | 0.013 | 0.011 | 0.219 | 0.915 | RNN | 0.008 | 0.007 | 0.149 | 0.925 | LSTM–GWO–FSO | 0.0055 | 0.0045 | 0.128 | 0.94 |
|
|