Research on Heating Load Forecasting Based on Support Vector Machine and Multiple Linear Regression

Various prediction models from the perspectives of short-term and long-term heat load prediction are summarized and verified for centralized heating heat load prediction. Firstly, historical data is preprocessed by removing abnormal data and standardizing the sample dataset. Secondly, support vector machine regression (SVM) model and multiple linear regression (MLR) are chosen for modeling and predictive analysis. Then, the mean square error (MSE) criteria are applied to evaluate the model performance by calculating the MSE values corresponding to different prediction results. Finally, historical production data of a heating company in Dalian City are used to verify the applicability of different prediction model.