文献类型：期刊文献

中文题名：基于ART2模型的信用风险评估

英文题名：Credit Risk Evaluation Based On ART2

作者：马海英[1]

机构：[1]华东理工大学商学院,上海200237

年份：2006

卷号：21

期号：4

起止页码：47

中文期刊名：华东理工大学学报（社会科学版）

外文期刊名：Journal of East China University of Science and Technology：Social Science Edition

收录：国家哲学社会科学学术期刊数据库

语种：中文

中文关键词：适应共振;神经网络;信用风险

外文关键词：ART2; neural network; credit risk

摘要：自适应共振模型是为了能够分类任意次序模拟输入模式而设计的,它可以按任意精度对输入的模拟观察矢量进行分类,较好地解决了前稳定性和灵活性问题,同时能够避免对网络先前所学的学习模式修改。本文将ART2模型应用于信用风险评估,通过实证比较研究,结果显示应用自适应共振模型进行信用风险评估在精度和准确性上,都优于其他神经网络模型和统计方法。
IART2 is designed to vector based on random precision, former study mode of the network. prior to other nerve network mode classify simulated input mode with random order. It can classify simulated observing which solves problems of pre-stability and flexibility; still, it can avoid amendment of This paper uses ART2 to evaluate credit risk, and its precision and accurateness are and statistical method through comparative research of real ease.