Short-term power load forecasting based on SVM

Short-term load is random and variable with many influence factors that are difficult to establish function relationship between power load. In this paper, the advantages and disadvantages of three commonly used short-term power load forecasting methods are compared based on the analysis of load characteristics, then a conclusion is drawn that support vector machine method, which can reflect several important characteristics of load and model the nonlinear relationship between load sequence and influence factors, is practical short-term for power load forecasting. Next, the steps for short-term power load forecasting are expatiated. Finally, A experiment of daily load forecasting of an area in Hubei is presented, which shows that the forecasting result based on support vector machine is better than those of others methods.