Font Size: a A A

A Study Of Personal Credit Scoring Methods Based On Genetic Programming Model

Posted on:2007-06-29Degree:MasterType:Thesis
Country:ChinaCandidate:Y Y JinFull Text:PDF
GTID:2189360212482349Subject:International Trade
Abstract/Summary:PDF Full Text Request
Credit evaluation is to test the ability to carry out the regulations and the degree of credit in marketing economic communications. In a mature marketing economic environment, credit is the foundation of a person and the"pass"to get into the society. After entering WTO, the strong credit sense of developed countries will have a big challenge to china which has dropped behind for more than 100 years in the development of personal credit scoring market. Personal credit is the key to the existence and development of our domestic commercial banks, while the personal credit evaluation system is still not complete, personal credit has become the"bottle-neck". Hence, the establishment of personal credit system is of great importance to our consumer market, and the adjustment of our economic structure and the debt structure as well.Based on consulting varieties of references, we proceed surrounding personal credit and its scoring technique. To begin with, we analyze former studies and conclusions of personal credit, combining the special conditions of our country, and present the variable sets that should be considered in domestic banks credit scoring as well, and screen out variable sets precisely and make some further process. Secondly, we introduce GP (Genetic Programming) theories systematically, then combine the theory with personal credit scoring and make practicality analysis for the whole model, and also demonstrate the reasoning process of the model in detail. In the end, under the condition of the above, we analyze the personal consumer credit data in a commercial bank in ShenZhen and check them up using GP model, and compare the total classification accuracies of four different typical credit scoring methods. As to GP model, it has a total classification accuracy of 92.8%. What's more, the running process is somewhat simple and we get precise result which is consistent with the fact. In conclusion, it is reasonable to consider using GP theory to establish personal credit scoring model in the internal of domestic commercial banks since personal credit scoring is blooming in our country nowadays.
Keywords/Search Tags:consumer credit, personal credit scoring, genetic programming model
PDF Full Text Request
Related items