SVM-based power consumer credit risk early warning method and system

Abstract

The invention relates to an SVM-based power consumer credit risk early warning method and system, and the method comprises the steps: setting a threshold value for a power consumer credit risk basic index, and dividing risk early warning intervals of different power consumer credit risk basic indexes according to the threshold value; selecting sample data, processing the sample data, distributingcorresponding weights to the power consumer credit risk basic indexes by using a principal component analysis method, calculating a risk early warning evaluation value of the sample data, and dividingrisk categories where the risk early warning evaluation value is located according to the risk early warning evaluation value; establishing a power consumer credit risk early warning SVM model, and predicting the power consumer credit risk; and taking the obtained optimal parameter value as a parameter value of a power consumer credit risk early warning SVM model, and optimizing the power consumer credit risk early warning SVM model. According to the method, parameter optimization is carried out through the SVM classification model, the prediction accuracy for the nonlinear relationship and the small sample model is high, and a good early warning effect is achieved in credit risk early warning.

Description

technical field [0001] The invention relates to the technical field of credit early warning, in particular to a method and system for early warning of credit risk of electric power users based on SVM. Background technique [0002] With the continuous deepening of power system reform, power companies are facing more and more credit risks for power customers. Power customers have the phenomenon of arrears of electricity bills, electricity theft, and illegal use of electricity, which affects the normal business activities of power supply companies, disrupts the normal order of electricity market transactions, and has a huge negative impact on the development of power companies and equipment updates. In order to prevent potential risks faced by power companies, credit risk assessment and early warning must be carried out for power customers. [0003] With the increasingly complex economic environment, the early warning methods of corporate credit risk are also constantly update...

AI Technical Summary

Problems solved by technology

[0012] Aiming at the deficiencies of the prior art, the present invention provides a SVM-based power user credit risk early warning method and system, aiming at the untrustworthy phenomenon that affects the transaction order existing in the power market, and solves the problem of early warning of the power user's purchase and sale of electric power risks in the power market

Images