Research Article
A Simple Method of Residential Electricity Load Forecasting by Improved Bayesian Neural Networks
Table 6
Results of BNN_16 model with different periods of historical data.
| Historic period | Training set (60%) | Test set (20%) | Computing time (s) | MSE | R | MAPE | MSE | R | MAPE | (%) | (%) |
| 1 month | 6.15e-2 | 9.61e-1 | 9.20 | 1.32e-1 | 9.14e-1 | 9.36 | 4 | 3 months | 9.94e-2 | 9.00e-1 | 11.73 | 1.54e-1 | 8.47e-1 | 11.92 | 4 | Half year | 7.61e-2 | 8.91e-1 | 10.88 | 8.88e-2 | 8.76e-1 | 11.05 | 8 | 9 months | 8.26e-2 | 8.91e-1 | 10.76 | 9.05e-2 | 8.83e-1 | 10.87 | 11 | 1 year | 8.28e-2 | 8.92e-1 | 10.63 | 8.89e-2 | 8.83e-1 | 10.59 | 12 |
|
|