Abstract: The fine scheduling of smart grids requires higher accuracy and robustness in bus load forecasting. Traditional methods, due to the fragmentation of multi-source data and insufficient model ...