Multiscale personalized federated load forecasting via enhanced ensemble empirical mode decomposition and Bayesian long short-term memory network

Key Points

• Enhanced ensemble empirical mode decomposition improves load forecasting accuracy significantly, especially in dynamic conditions.
• Bayesian long short-term memory networks provide a robust method for time series analysis in load forecasting applications.
• Observational analysis across multiple scales enables personalization of forecasting models tailored to specific regional energy demands.
• These findings may enable more efficient energy management strategies, promoting resource optimization across varying load profiles.