Credit default prediction method based on svm under differential privacy

Abstract

The invention discloses a credit default prediction method based on SVM under differential privacy in the technical field of credit default, including the following steps: S1: data preprocessing; S2: variable selection; S3: according to the nature of differential privacy serial combination, design differential privacy weighting SVM optimization model, the present invention provides an effective solution to the differential privacy SVM learning problem under data imbalance, and the solution can specifically solve the data imbalance problem faced when using differential privacy SVM learning to predict customer default, and is suitable for Application scenarios with data imbalance including credit card default prediction, such as disaster prediction, medical diagnosis and other fields are also applicable to the technical solution of the present invention.

Description

Credit default prediction method based on SVM under differential privacy technical field [0001] The present invention relates to the technical field of credit default, in particular to a SVM-based credit default prediction method under differential privacy. Background technique [0002] With the rapid development of social economy, more and more people use credit cards to achieve advanced consumption. in people enjoying the letter With the convenience of consumption brought by cards, more and more credit card debt problems also arise. Some people are unable to Timely repayment of credit card debts leads to credit card defaults, causing financial institutions and consumers to suffer huge economic losses and serious damages. The existing financial order is broken, and consumer financial information is hit. How to effectively identify potential credit card default customers and reduce credit The phenomenon of card default is a huge challenge faced by financial institut...

AI Technical Summary

Problems solved by technology

And in the context of differential privacy, the existing theory has the defect of ε-differential privacy insatiable when solving such problems, and cannot be directly applied to solve the problem of data imbalance through simple theoretical expansion.

This technical defect is a phased problem faced in the development of differential privacy SVM learning technology from theoretical research to practical application.

Images