Short Term Load Forecasting Model Based on Kernel-Support Vector Regression with Social Spider Optimization Algorithm

چکیده مقاله

Short-term load forecasting in power system is an important factor planning and electricity marketing. Due to the uncertainty of the load demand, many studies have been devised for nonlinear prediction methods. In this paper, a hybrid approach consisting of support vector regression (SVR) and social spider optimization (SSO) is proposed for short term load forecasting. The SVR technique has proven to be useful in nonlinear forecasting problems. To improve accuracy of SVR parameters are tuned using SSO. The SSO algorithm is based on the simulation of cooperative behavior of social-spiders and helps in achieving good results.