Font Size: a A A

Application Research Of Data Mining Techniques In SME Credit Risk Management

Posted on:2014-12-26Degree:MasterType:Thesis
Country:ChinaCandidate:H L WangFull Text:PDF
GTID:2268330425980519Subject:Computer application technology
Abstract/Summary:
Small and medium enterprises (SMEs) play indispensable role in thedevelopment of national economy. SMEs are an important vigor in China’s nationaleconomic and social development, and promote the development of small and medium-sized enterprises are important foundation to maintain steady and rapid economicdevelopment. And it’s a major strategic task of the people’s livelihood and socialstability. But due to their small economies of scale, obvious impact of macroeconomicdevelopment, the imperfection of the credit rating system, the lack of technical meansof financial institutions ratings and many other factors, all these result in corporatefinance difficult.On the basis of in-depth analysis of the constitution of the small and mediumenterprises’ credit risk, combined with China Construction Bank, the credit rating ofsmall business customers (Trial)(built in total Fa [2009] No.101), the thesis select sixfirst class indexes including the quality of the operators, strength, solvency, businessgrowth, and my line relationship, social responsibility as the main parameters of theevaluation of small and medium enterprises credit rating, which consists of21secondary indexes. All these indexes form the SMEs’ credit rating model. Withdata mining technology, the thesis put forward reasoning algorithm by using fuzzy rule,developing rating system for small and medium enterprises, and rating the credit forSMEs.The rating system runs in bank’s local network, so C/S framework is more safeand convenient. The Database server run environment is Microsoft2003Server+SQL Server2005, and the client runs on Windows XP/SP2or upper platform. Thewhole system contains basic data input&output, credit rating, rating historymaintenance, credit indexes purify.Finally, the thesis describes the systematic technical implementation and theexperimental process. The experimental data show that the proposed fuzzy rule reasoning algorithm can effectively solve the problem of non-standard data, thecomplex and variety of the indicators during the small and medium enterprises’ creditrating process.
Keywords/Search Tags:Small&medium enterprise, Credit rating, Data mining, Fuzzy rule
Related items