Research On Personal Credit Evaluation Based On Enhanced Technology Of XGBoost Multi-Model Fusion

Under the background of the request of the central government to build a social credit system,local governments at all levels are gradually establishing and improving the social credit system,innovating financial credit products,improving financial services,safeguarding the personal information security of financial consumers,and protecting the legitimate rights and interests of financial consumers.The construction of social credit system makes many financial enterprises get a perfect upgrade of their work business.With the maturity and rapid development of mobile interconnection,big data,cloud computing and other technologies,the use of large data and Internet technology,through the establishment of multi-dimensional data model for customers to issue small loans online products,greatly improving enterprise efficiency and customer experience.The current credit data information is huge and complex,which needs to be observed from the perspective of large data,evaluated by more advanced data algorithms,and analyzed with the current social situation.Accurate evaluation of personal credit rating and improvement of local data are two main contents of model research.Based on these two contents,and on the basis of systematic research on various basic models,under the guidance of multi-model fusion theory,this paper proposes the models of multi-model fusion technology based on XGBoost and multi-model fusion technology enhancement.In the process of personal credit data processing,we use data missing value filling and Pearson coefficient correlation analysis and other methods.Then,XGBoost algorithm is used to classify and predict data,and multi-model fusion technology is used to improve the accuracy of classification and prediction.In the analysis of the factors affecting personal credit,the interaction between the factors is constantly analyzed,and the influence degree of the main factors on personal credit is discussed.The evaluation method we established has high applicability.Supported by data conditions,the classification and prediction effect of personal credit data is excellent.This series of research lays a solid foundation for the establishment of credit supervision system.