Countries all over the world have similar goals and aspirations to improve their economic stability. To this end, they must combat the root causes, and among the many problems they have in common are money launderers and fraudsters, who pose a serious threat to sustained economic growth. In this sense, insurance fraud, especially automobile insurance fraud, is a common fraud topic which causes significant financial losses for insurance companies.

Dishonest claims impose a significant financial strain on insurance companies which influence on the entire industry. Because of this, insurers are constantly seeking for more effective detection systems to get beyond the limitations of traditional techniques; therefore, building a reliable and effective fraud detection model is crucial to maintaining the insurance providers' financial stability and reputation.

This capstone project's major objective is to use machine learning techniques to develop a reliable system for detecting car insurance fraud. This study intends to examine multiple supervised machine learning algorithms using classification method, and  assess their efficacy, to ultimately select the most accurate model with significant accuracy for identifying fraudulent insurance claims.