Social factor is considered for the linear regression models besides weather file.
Simultaneously optimize all the coefficients for linear regression models.
SARIMA combined with linear regression is used to forecast the heat demand.
The accuracy for both linear regression and time series models are evaluated.