Dealing with abnormal days is the key in the short-term load forecasting. To integrate the advantages of several forecasting models and improve the accuracy of load forecasting results

based on data mining techniques

a decision-tree combined shortterm load forecasting model is proposed to mainly solve the weathersensitive factors’ influence on the power load. The statistic analysis shows that the accuracy of the short-term load forecasting in abnormal days has greatly increased by using the improved strategy. Meanwhile

the actual forecast results of Anhui Province’s total electric power loads validate the effectiveness and the superiority of the method compared with the original system.