Short-Term Load Forecasting Based on Attribute Reduction Algorithm of Rough Sets and Support Vector Machine

Abstract

A short-term load forecasting model based on two integrated intelligent algorithms, i.e., attribute reduction algorithm of rough sets and support vector machines (SVM), is proposed. At first, according to historical data a attribute decision table is built up and the data mining is performed by means of attribute reduction algorithm, thus the kernel factors influencing loads is determined and using them as the input vectors of SVM the load forecasting is conducted. Forecasting results of calculation examples show that comparing with traditional SVM model that chooses input vectors in the light of experience the forecasting accuracy is evidently improved and is more suitable to short-term load forecasting.