Assessing The Efficacy of Linear Regression and Long-Short Term Memory in Refining the Accuracy of Energy Consumption Forecasts for The Purpose of Optimized Energy Distribution

The primary aim of this research is to assess the efficacy of Linear Regression (LR) and Long-Short Term Memory (LSTM) models in refining the accuracy of energy consumption forecasts for optimized energy distribution in modern electricity systems. Materials And Methods: A comprehensive dataset containing historical energy consumption data is meticulously compiled and preprocessed. LR and LSTM models are developed, trained, and assessed using standard evaluation metrics, with a particular emphasis on Mean Absolute Error (MAE) and Root Mean Square Error (RMSE). Results: Comparative analysis reveals that while LR model exhibits an MAE of 0.732 and an RMSE of 1.512, the LSTM model showcases improved performance with an MAE of 0.621 and an RMSE of 1.305. These results underscore a discernible disparity between the two models, with the LSTM model demonstrating superior accuracy. Statistical analysis further corroborates these findings, with a calculated p-value of 0.017 (p < 0.05) for both accuracy and loss, indicating a statistically significant advantage in favor of the LSTM model. Conclusion: In conclusion, this study establishes that LSTM models hold promise in enhancing the precision of energy consumption forecasts, thereby facilitating optimized energy distribution and bolstering energy efficiency in modern electricity systems. By leveraging advanced machine learning techniques, such as LSTM, significant strides can be made towards achieving sustainable energy management practices and maximizing energy efficiency in the contemporary energy landscape.