Machine Learning Algorithms for Auto Insurance Fraud detection: Comparative Study
Prashant Gupta, Risabh Bhatnagar, Shalu Tyagi, Achin Jain, Arvind Panwar, Arun Kumar Dubey
Abstract
Insurance fraud is a massive burden for the industry and causes billions in economic damage yearly. This research deals with the difficulties in fraud detection of insurance companies, focusing on machine learning approaches. Based on factual data from an insurance company the study kind use SVM, Decision Tree model, Random Forest Ada Boost Gradient Boosting and Stochastic Gradient Boosting, XGBoosts Extra Trees in this case. The assessment of this models is through various proper analysis which include confusion matrix and main performance metrics like accuracy, precision, recall, F1 and so on. This study aims to provide insights that would be beneficial to improve the fraud detection strategy in insurance industry by comparing the performance of models used.