Credit assessment in the electricity market by least squares support vector machines

Credit assessment is crucial for the marketing of power distribution enterprises in the electricity market. But credit assessment on the power clients belongs to typical multi-classification and is still unsolved, due to the small-sampled problem in the market. So this work aims at proposing a novel credit assessment model of the electric power consumers based on least squares support vector machines (LS-SVM). In the proposed work, multi-pattern identification of consumer credits is accomplished by LS-SVM that builds the nonlinear mapping of the credit indexes and the corresponding scores implemented by the linear mapping in the high-dimensional feature space according to statistical learning theory. In this way, credit assessment is solved by this special kernel technology to improve the classifiable abilities of the samples. Case studies are carried out to test the proposed model.