Research Article
Hybrid Machine Learning Model for Electricity Consumption Prediction Using Random Forest and Artificial Neural Networks
Table 6
Performance of proposed predictive models on the test sample.
| Predictive model | No. of nodes in input layer | RMSE | MAPE | NRMSE | SMAPE | R2 | Acc |
| BPNN | 21 | 0.0415 | 8.1999 | 0.0458 | 21.0117 | 0.9638 | 0.9631 | PCA 1 + BPNN | 9 | 0.0601 | 10.5201 | 0.0596 | 28.0458 | 0.6910 | 0.7921 | PCA 2 + BPNN | 13 | 0.0405 | 8.5201 | 0.0499 | 22.0321 | 0.8913 | 0.8908 | SWR + BPNN | 12 | 0.0454 | 8.1921 | 0.0488 | 21.0112 | 0.9492 | 0.9487 | RF 1 + BPNN | 18 | 0.0416 | 8.1992 | 0.0470 | 21.0114 | 0.9932 | 0.9926 | RF 2 + BPNN | 11 | 0.0419 | 8.1994 | 0.0475 | 21.0117 | 0.9930 | 0.9923 |
|
|