Classifiers based on Developed SVM for Credit Analysis in Electricity Market

The credit analysis on electricity users is of great use for the decision-making and management of market trades in the market. But due to the short history and the complexity of client information, the credit analysis in electricity market has indeed formed a typical nonlinear classification problem with small samples, which is still unsolved by traditional methods. To solve above problem, a novel classifiers based on developed SVM for the credit analysis in electricity market is presented, where two algorithms are integrated: 1)support vector machine (SVM) is the basic algorithm with special adaptability and advantage in nonlinear higher-dimensional pattern identification with small samples; 2) independent component analysis (ICA) is an excellent tool for blind signal separation. In the model, first the attributes of credit analysis for electricity uses are reconstructed by ICA in order to overcome the degradation of the latent noise and redundancy in SVM inputs. Second, the mined attributes with better information are fed to SVM for credit classification. In this way, the accuracy of SVM classifier is fatherly enhanced by combining with ICA. And then the performance of the credit analysis is improved. Simulation result shows that the proposed method may increase the accuracy of credit analysis.