Medium And Short Term Electric Power Load Forecasting Based On Support Vector Machine

Medium and short term load forecasting has become increasingly important since the competition of the electric power market is more drastic and has gradually become one of the major areas of research in recent years. This paper introduced Support Vector Machine (SVM) and Particle Swarm Optimization (PSO) theories. This paper also introduced methods of data preprocessing and analyzed the data to identify its internal laws. After that this paper gained the training data respectively based on the original and pre-processed data and used Sequential Minimal Optimization (SMO) arithmetic to achieve the fast training of the Support Vector Machines (SVM). Finally the load forecasting result was gained. Then this paper did load forecasting with the data by classified according to week property. The simulation results show that we can get high forecasting accuracy by using the pre-processed data and the higher forecasting accuracy all over this paper by classification forecasting according to week property.