| With the rapid development of the credit economy,the application of "credit consumption" in the country's economic operations has changed significantly."How to make a balance between credit lending and credit default rates,and how to reduce default rates" is one of the main technical issues.By analysing the basic social attributes of individuals and credit-related information,and using machine learning methods to build models to quantify credit risk,a suitable equilibrium can be found in the problem of individual credit rationing,and it can provide an effective solution to reduce credit risk.While the international economy is in the doldrums,has the epidemic had an impact on credit operations? Has it had a significant impact? How should the impact of the epidemic be addressed? To address these questions,this paper selects a desensitised dataset of personal credit for 2019-2020 from the official website of Lending Club in the US for analysis.Due to the large number of fields and information interference in the original dataset,pre-processing work such as missing value addition and unbalanced data processing was done on the original data before doing the risk prediction model on the dataset.In the process of fitting the model,the data set was again considered to be divided into two sets of data before and after the epidemic,and a comparative analysis of the three sets of data was conducted to explore whether the epidemic had an impact on credit risk.The final conclusion is that the risk of default in personal credit business has increased due to the epidemic,and the data before and after the epidemic have changed to a certain extent,the effect of the model fitted to the overall data is inferior to that of the model fitted to the grouped data,which does not reflect well the changes caused by the epidemic.Finally,the model predicts the risk of creditors with intermediate repayment status,and makes recommendations for the development of credit business based on the prediction results. |
PDF Full Text Request