<![CDATA[This paper presents the effectiveness of the ensemble empirical mode decomposition-long short-term memory (EEMD-LSTM) model for short term load prediction. The prediction performance of the proposed model is compared to that of three other models (LR, ANN, LSTM). The contribution of this research lay in developing a novel approach that combined the EEMD-LSTM model to enhance the capability of industrial load forecasting. This was a field where there had been limited proposals for improvement, as these hybrid models had primarily been developed for other industries such as solar power, wind power, CO2 emissions, and had not been widely applied in industrial load forecasting before. First, the raw data was preprocessed using the IQR method, serving as the input for all four models. Second, the processed data was then used to train the four models. The performance of each model was evaluated using regression-based metrics such as mean absolute error (MAE) and mean squared error (MSE) to assess their respective output. The effectiveness of the EEMD-LSTM model was evaluated using Seojin industrial load data in Vietnam, and the results showed that it outperformed other models in terms of RMSE, n-RMSE, and MAPE errors for both 1-step and 24-step forecasting. This highlighted the model's capability to capture intricate and nonlinear patterns in electricity load data. The study underscored the significance of selecting a suitable model for electricity load forecasting and concluded that the EEMD-LSTM model was a dependable and precise approach for predicting future electricity assets. The model's robust performance and accurate forecasts showcased its potential in assisting decision-making processes in the energy sector.]]>
Copyrights © 2023
