Discrimination of Insurance Fraud Based on Machine Learning

Abstract

In the insurance industry, Insurance fraud is a common phenomenon. However, according to statistics, among all types of insurance, automobile insurance fraud is the high incidence time of Insurance fraud. Based on 39 characteristic variables in the insurance claims database, this paper completes data preprocessing through normalization and coding of Categorical variable; Then it analyzes the correlation between data characteristics and Insurance fraud; Then, the principal component analysis method is used to reduce dimensionality and extract features from multi-dimensional features; Finally, the SVM classifier is trained to effectively identify Insurance fraud. The research results show that the model is effective in identifying Insurance fraud, and can achieve 79% accurate discrimination.