Portfolio Details
Project information
- Category: Fraud Detection
- Client: Personal
- Project date: 6 April 2025
- Project URL: -
Description
This project focuses on analyzing insurance claim data to identify patterns of fraudulent behavior and build predictive models for early fraud detection. Through exploratory data analysis, clustering, and machine learning models (Logistic Regression, Decision Tree, Random Forest), the project highlights significant fraud indicators such as gender, accident area, claim size, and vehicle attributes. The Random Forest model proved to be the most stable and accurate, making it suitable for implementation in an early warning fraud detection system.