Electrical energy is a basic human need. With the increase in vulnerable populations and technological developments, the demand for electrical energy is also increasing. Forecasting or predicting electricity usage obtain a necessity for planning future energy resources. Electricity usage accurately forecasting is imperative and challenging. However, electricity usage data in society contains significant, nonlinear, and complex variances. The neural networks effectively overcome complex and nonlinear data problems, while bagging can reduce substantial conflicts. Therefore, this study combines a neural network with bagging to improve the prediction performance of electricity consumption. The results show that both data sets significantly reduce RMSE values between neural network applications using the neural networks and bagging techniques. The comparison between the neural network and bagging shows that the RMSE value is smaller than that of the neural network. Based on these results, the combination of neural networks and bagging performs better than a particular neural network.
| Author | Tyas Setiyorini; Frieyadie; Andrianingsih; Maryanah Safitri; Tati Mardiana; Mari Rahmawati |
| Conference Type | Scopus indexed Reputable International Conferences |
| Published in | 2022 International Conference on Information Technology Research and Innovation (ICITRI) |
| Quartile Rank | Non-Q |
| DOI | 10.1109/ICITRI56423.2022.9970204 |
| Publisher | IEEE |
| Pages | 107-111 |
| Semester | Ganjil 2022/2023 |
| URL | https://ieeexplore.ieee.org/document/9970204 |
| Download/Mirror | Download |