Evaluation of Electricity Customers Credit Based on Genetic Algorithm and Neural Network

摘要

The question that electricity charge is in arrears becomes increasingly serious, so power supply enterprises and the entire society pay more attention to credit evaluation of power clients. Basing on the analysis of factors influencing the credit of power clients, index system suitable for credit evaluation of power clients is established. After analyzing the existing methods of power clients' credit evaluation, model of power clients' credit evaluation is established by BP neural network. In this model, the connection weight and the valve value of BP neural network are optimized by genetic algorithm. And because genetic algorithm is good at the global search, so combined with BP neural network, the existing two problems that BP neural network will easily fall into local minima and the convergence rate of BP neural network is slow are solved. The example research of power clients of Baoding power supply enterprise indicates that errors are much smaller than that calculated by BP neural network and the evaluation result of genetic neural network is satisfied.