Retail Baking Credit Scoring Model Development And Implemetation

With the rapid development of China’s economy, the domestic banking industry is gradually connecting with the world. In recent years, along with the new Basel capital accord will gradually be introduced in, the domestic most Banks are gradually began to the new Basel capital accord guidance for credit risk measurement, risk quantification are also gradually become the main credit risk management method. Because our country credit information collection and the corresponding system construction work starts relatively late, data quality and data length of time are difficult to meet the needs of the development of the quantitative model, so the current methods of risk identification of China’s banking industry is mainly in combination with quantitative and qualitative methods. The main purpose of this paper is how to develop quantitative and qualitative methods for personal credit scoring model.Based on the practical project of a commercial bank, will be housing loans as an example, elaborated how to regression algorithm through data mining logic, combined with the whole process of personal credit scoring model analytic development of quantitative and qualitative method combining the. At the same time also on the model used in real bank credit business, given the corresponding recommendations. In order to be able to provide help to the quantitative management of credit risk of bank.Personal credit scoring model established in this project, is based on the quantitative model of historical data. With the experience of experts to analyze the adjustment, qualitative modeling method is introduced, and the results of qualitative analysis to quantitative models were modified, eventually reached the adjustment result. In order to make up for lack of data, thus achieve the purpose of optimization model.Through data collection, extraction, pretreatment, model definition, subdivision, development, verification steps, establishes individual credit scoring model, and the model is put to use. From the model used across time for validation sample results, the developed model, predictive ability in individual housing loans in the relatively high, basically reached the expected goal. On this basis, we propose the phase information of commercial banks on the one hand to the collection, and gradually into the model, which will further increase the prediction ability of the model. On the other hand, also need to be analyzed from the angle of market risk, to ensure that the risk control ability.