文章基本信息

标题：IMPACT OF THE SEGMENTATION FOR MODELS OF ANALYSIS IN GRANTING OF CREDIT: A COMPARISON OF RESULTS
本地全文：下载
作者：Vanessa Anelli Borges ; Fabiano Guasti Lima ; Tabajara Pimenta Júnior 等
期刊名称：Business and Management Review
印刷版ISSN：2047-0398
电子版ISSN：2047-0398
出版年度：2014
卷号：3
期号：10
页码：01-16
出版社：Global Research Society
摘要：This paper explores the contribution of segmentation data for analysis models in granting of credit. The decision of granting credit directly impacts business continuance, which justifies the importance of managing the risk of default. Given a base of 50,000 customers, it’s applied discriminant analysis to classify the predictive ability of this technique. Neural Networks are also used with this goal. Further, the data is segmented by cluster analysis K-Means, applies discriminant analysis, and then the Neural networks for each cluster formed. The sum of hits obtained in the second step is compared with the previous results. The process is repeated for the TwoStep Cluster analysis. The classifying capacity is higher for the two techniques combined with the K-means analysis. The results obtained with the TwoStep Cluster analysis weren’t satisfactory. Finally, it was found that Neural networks have better classification ability
关键词：neural networks; discriminant analysis; cluster analysis; credit risk.