BUILDING A CREDIT SCORING MODEL BASED ON THE TYPE OF TARGET VARIABLE

With the rapid development of big data and Internet technologies, companies engaged in big data financial platforms collect and systematize massive data through their own platforms, improve credit scoring parameters and use machine learning methods to conduct complex and scientific credit scoring assessments. Thus, banks face big problems when building credit scoring. Based on the limitations of the existing system and methods of personal credit rating, it is necessary to study personal credit rating based on machine learning methods, improve the parameters and scoring system of personal credit rating, clarify data collection channels, use dynamic desensitization technology. To reduce the sensitivity of the data, the LOF test method is used to verify the emission data and the random forest method is used to fill in missing data values. Then you use the gradient-boosting decision tree method to view important indicators, process proven indicators using a metric system model based on logistic regression, and get a personal credit score. Finally, the model is tested using a BP neural network, and the model is used to predict the level of personal credit. The study shows that machine learning can further improve the accuracy of individuals' credit ratings and provide a scientific basis and background information for commercial banks' credit ratings. 

1. Sorokin, A.S. Primenenie zakonov raspredelenija sluchajnyh velichin dlja modelirovanija jekonomicheskih javlenij i processov [Tekst] : monografija. / N.Ja. Bambaeva, A.C. Sorokin – M.: MJeSI, 2010. – 156 c. (in Russian).

2. Liao Biying. Research on Personal Credit Rating Model of Commercial Banks [D]. Nanchang: Jiangxi University of Finance and Economics. – 2016. (in Russian).

3. Yu Ya. Comparative Research on Personal Credit Evaluation of Zhima Credit Company [D]. Changsha: Hunan University. – 2017. (in Russian).

4. Chen Zhipeng. Comparative analysis of personal credit information management systems at home and abroad [D]. Harbin: Heilongjiang University. – 2017. (in Russian).

5. Sorokin, A.S. K voprosu ocenki soglasovannosti mnenij jekspertov pri ispol'zovanii metodov jekspertnogo ocenivanija v kreditnom skoringe. [Tekst] /A.C. Sorokin // Rol' biznesa v transformacii obshhestva – 2014: Sb. st. po mat. IX mezhdunar. nauchn. kongr. – M.: «Jeditus». – 2014. – s. 281-283. (in Russian).

6. Cao Jie, Shao Xiaoxiao. Research on Personal Credit Evaluation Model Based on Information Gain and Bagging Integrated Learning Algorithm [J]. Practice and Understanding of Mathematics. – 2016, 46 (8): 90-98. (in Russian).

7. Tan Zhongming, Xie Kun, Peng Yaopeng. Research on Credit Risk Evaluation of P2P Online Loan Borrowers Based on Gradient Boosting Decision Tree Model [J]. Soft Science. – 2018, 32 (12): 136-140. (in Russian).