摘要
提出了一种基于遗传算法改进的神经网络企业信用评级模型.利用BP神经网络自适应和自学习的特点,通过遗传算法实现对神经网络连接权重和阈值的修正与优化,在一定程度上解决了BP神经网络存在收敛速度较慢和可能落入局部最小点的问题.实验表明,该模型在企业信用评级上有着较高的准确率,具有很好的应用价值.
In order to meet the requirements of the corporate credit rating, based on neural networks and enterprise credit management research,corporate credit rating model based on BP neural networks is proposed. Using the BP neural network adaptive and self-learning characteristics, by using genetic algorithm of neural network connection weights and thresholds correction and optimization, solve the artificial neural network convergence speed and might fall into the local minimum points. Proved by experiments,the model has a higher accuracy rate on corporate credit rating.
出处
《东北师大学报(自然科学版)》
CAS
CSCD
北大核心
2013年第3期62-68,共7页
Journal of Northeast Normal University(Natural Science Edition)
基金
国家"863"计划基金资助项目(2008AA040205)
吉林省教育厅"十二五"科学技术研究基金资助项目(吉教科合字2012第404号)
关键词
信用评级
BP神经网络
遗传算法
权重
credit rating
BP neural network
genetic algorithm^weight