数据挖掘在商业银行信用风险管理中的应用研究.pdfVIP

电子科技大学硕士论文 文提出采用训练样本和预测样本（均含正常客户和违约客户）配对， 通过模型训练，再应用模型对预测样本进行预测，分析准确率。4 ） 本文是应用数据挖掘对信用风险进行分析的一个较为全面的缩影， 涵盖了数据挖掘的重要方面如：数据抽取、清洗、转换、模型化、 评价等，对信用风险的其他方面的研究有重要的借鉴意义。 最后本文总结研究成果，指出研究的不足，并提出进一步开展 研究的方向及领域。 关键字 金融风险/信用风险/数据挖掘/决策树/神经网络 ii 电子科技大学硕士论文 THE APPLIED STUDY ON DATA MINING BASED CREDIT RISK MANAGEMENT OF COMMERCIAL BANK ABSTRACT The study on the credit risk management of commercial bank is major studying topic in financial area. The paper takes the applied study on data mining based credit risk management as studying objective, applies the method combining with theoretic study and practical study. The theoretic study forms the basic viewpoints and models of financial risk management, commercial bank credit risk and data mining mainly by acquiring the newly applying studying results about financial risk management, commercial bank credit risk and data mining, packing up and mining the relative literatures. The practical study takes the quantitative study and comparative study according to the credit card system of Shanxi branch of one national bank. In accordance with credit data, the thesis adopts comparative study through training samples and predictive samples, normal clients and default clients, four predictive models. First of all, for theoretic study, the paper acquires the viewpoints as follow: 1) Financial risk is the most concentrative, potential destructive risk in social life; it can not be eliminated, but can be controlled and must be controlled. 2) Credit risk is the most important risk in commercial bank; in west countries, modern mature models a