Logistic-BP神经网络的变权组合模型在公司信用评级中的应用

对公司建立信用评价模型,以便及早发现信用危机信号,不仅有利于公司的经营管理,也有利于投资人的投资决策。采用logistic回归模型和BP神经网络相结合的变权组合评价方法,试图寻找出降低公司信用风险的有效措施。试验结果表明,变权组合预测模型不仅有较高的预测精度,同时也有较好的稳定性,变权组合模型优于单一模型。

To Apply on Variable Weight Combination Model of the Logistic-BP Neural Network in Company Credit Rating

Establishment of companies credit evaluation model,so as to discovery credit crisis signal,is not only beneficial to management,but also to investment decision .The combination of logistic regression model and BP Neural Network is adopted in the paper. By testing,the result shows that the combination of these two methods not only have high prediction precision,but also have better stability. The variable weight combination model of logistic and BP is better than single model.