Research On The Construction Of Personal Credit Scoring Model For Small And Medium-Sized Commercial Banks

Since 2018,the asset quality of small and medium-sized commercial Banks has shown a serious downward trend.By the end of September 2018,the non-performing loan ratio of urban commercial Banks and rural commercial Banks has reached 1.67% and 4.23% respectively.The non-performing loan ratio of urban commercial Banks has increased slightly year-on-year,while that of rural commercial Banks has reached an alarming level.Compared with these small and medium-sized commercial Banks,the non-performing ratio of state-owned Banks and national joint-stock commercial Banks remained basically unchanged or even showed a downward trend.The main reason is that state-owned Banks and joint-stock Banks with nationwide business can attract better customers,and they can better disperse risks due to their wide business coverage.What's more important is that state-owned Banks and joint-stock Banks,due to their early start,sound rules and regulations,and good risk control system and access mechanism,can still operate in a stable and orderly manner under the overall economic downturn.In contrast,small and medium-sized commercial Banks have relatively poor customer groups and no good investment projects.Coupled with the intervention of the local government,it is easy to generate greater risks when the overall economy fluctuates.For small and medium-sized commercial Banks,the risk tolerance is relatively weak,and most of their customers are light individual customers,so it is very necessary to establish their own personal credit score system.The overall scale of small and medium-sized commercial Banks is small,and there is a big gap between the personnel allocation system and state-owned Banks and large joint-stock Banks,which determines the unsoundness of the whole system.At present,most of the research in China is about algorithms,and the role of algorithms cannot be ignored.However,for the establishment of the whole bank credit score model,algorithms alone are not enough.The data dimensions of Banks are large and relatively intersecting.Most of the algorithms are based on the exclusion of collinearity of multi-dimensional data.However,there are still incomplete data in Banks.Based on my internship experience in a bank and the research results of scholars at home and abroad,this paper establishes a credit score model system for small and medium-sized commercial Banks.Finally,the paper analyzes the way of judging the fitting degree of the credit scoring model and the development direction of the credit scoring model construction of small and medium-sized commercial Banks.