Litcius/Paper detail

LSTM vs. Prophet: Achieving Superior Accuracy in Dynamic Electricity Demand Forecasting

Saleh Albahli

2025Energies24 citationsDOIOpen Access PDF

Abstract

Accurate electricity demand forecasting is critical for improving energy efficiency, maintaining grid stability, reducing operational costs, and promoting sustainability. This study presents a novel hybrid forecasting model that integrates Long Short-Term Memory (LSTM) networks and Prophet models, leveraging their complementary strengths through a dynamic weighted ensemble methodology. The LSTM component captures nonlinear dependencies and long-term temporal patterns, while Prophet models seasonal trends and event-driven fluctuations. The hybrid model was evaluated using a comprehensive dataset of hourly electricity consumption from Ontario, Canada, achieving a Root Mean Square Error (RMSE) of 65.34, Mean Absolute Percentage Error (MAPE) of 7.3%, and an R2 of 0.98. These results demonstrate significant improvements over standalone LSTM, Prophet, and other State-of-the-Art methods, highlighting the hybrid model’s adaptability and superior accuracy. This study underscores the practical implications of the hybrid approach, particularly in energy grid management and resource optimization, setting a new benchmark for time series forecasting in the energy sector.

Topics & Concepts

Mean absolute percentage errorBenchmark (surveying)Mean squared errorComputer scienceAdaptabilityElectricityGridArtificial intelligenceArtificial neural networkStatisticsEngineeringMathematicsEconomicsManagementGeometryGeographyGeodesyElectrical engineeringEnergy Load and Power ForecastingForecasting Techniques and ApplicationsGrey System Theory Applications