TY - GEN
T1 - Comparison of Electricity Usage Forecasting Model Evaluation Based on Historical Load Dataset Duration Using Long Short-Term Memory Architecture
AU - Salleh, Nur Shakirah Md
AU - Suliman, Azizah
AU - Jørgensen, Bo Nørregaard
N1 - Funding Information:
The publication of this paper was funded by URND TNB Seeding Fund: U-TE-RD-20-08. The authors would like to thank the Institute of Informatics and Computing in Energy (IICE), Universiti Tenaga Nasional (UNITEN) for providing a platform to collaborate with the Center for Energy Informatics, Southern Denmark University (SDU).
Publisher Copyright:
© 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
PY - 2022
Y1 - 2022
N2 - Electricity prediction helps electric power companies to generate sufficient electrical power to consumers. The primary source used in performing forecasting is historical electricity usage. This research identified the optimum historical load data period in generating the best model for short-term forecasting of a household. The experiment applied Long Short-Term Memory (LSTM) architecture using Adaptive Learning Rate Method (Adadelta) on four categories of dataset: one-year, two-years, three-years, and four-years. The models produced were evaluated using mean squared error (MSE) and mean absolute error (MAE). The model generated from two-years of historical data performed the best among all other models with MSE value of 0.133621 and MAE value of 0.050653. The experiment was enclosed with the application of the model to predict the electricity usage of the following year, shown in two sample categories: one day and one week. Then, the prediction results were compared with the actual load.
AB - Electricity prediction helps electric power companies to generate sufficient electrical power to consumers. The primary source used in performing forecasting is historical electricity usage. This research identified the optimum historical load data period in generating the best model for short-term forecasting of a household. The experiment applied Long Short-Term Memory (LSTM) architecture using Adaptive Learning Rate Method (Adadelta) on four categories of dataset: one-year, two-years, three-years, and four-years. The models produced were evaluated using mean squared error (MSE) and mean absolute error (MAE). The model generated from two-years of historical data performed the best among all other models with MSE value of 0.133621 and MAE value of 0.050653. The experiment was enclosed with the application of the model to predict the electricity usage of the following year, shown in two sample categories: one day and one week. Then, the prediction results were compared with the actual load.
KW - Electricity
KW - LSTM
KW - Machine learning
KW - Prediction
U2 - 10.1007/978-981-16-8515-6_51
DO - 10.1007/978-981-16-8515-6_51
M3 - Article in proceedings
AN - SCOPUS:85127698632
SN - 9789811685149
T3 - Lecture Notes in Electrical Engineering
SP - 675
EP - 686
BT - Proceedings of the 8th International Conference on Computational Science and Technology - ICCST 2021
A2 - Alfred, Rayner
A2 - Lim, Yuto
PB - Springer
T2 - 8th International Conference on Computational Science and Technology, ICCST 2021
Y2 - 28 August 2021 through 29 August 2021
ER -