Short-Term Electricity Load Forecasting Based on ICA and LSSVM

Abstract

Short-term electricity load forecasting is a difficult work because the accuracy of forecasting is influenced by many unpredicted factors whose relationships are commonly complex, implicit and nonlinear. By studying the methods proposed by other scholars, a mew method, ICA (independent component analysis) -LSSVM (least squares support vector machine) is proposed by this paper. The first step of this method is to apply ICA to LSSVM for feature extraction. By using ICA, the original higher dimensional inputs will be transformed into other lower dimensional features. These new features are then used as the inputs of LSSVM to solve the load forecasting problem. By learning and training, we use the data of this subset to get the solution and find interrelationship of input and output by the LSSVM. This method has the advantage of self adaptability to weaken the human factors in fixing weight index and better extensive capability than traditional methods. It also has better convergence ability and strong global search ability, which consumes less time than other methods. Practical examples are cited in this paper to illustrate the process. The ICA-LSSVM method can also be used in other forecasting problem.