VictorUdeogu/ai-claims-fraud-detection

This project demonstrates how AI and data analytics can be used to detect fraudulent insurance claims — inspired by real-world processes from InsurTech platforms operating across Africa. # ai-claims-fraud-detection This project demonstrates how AI and data analytics can be used to detect fraudulent insurance claims — inspired by real-world processes from InsurTech platforms operating across Africa. ## Overview The model analyses synthetic claims data to identify patterns of potential fraud, delayed reporting, or inconsistent entries. It mimics Curacel’s approach to using AI to improve claims processing and reduce financial losses. ## Dataset - Synthetic dataset generated in Python (10,000+ sample claims) - Key columns: claim_id, amount, claim_type, days_to_report, fraud_flag ## Methods - Exploratory Data Analysis (EDA) - Logistic Regression for fraud classification - Random Forest for anomaly detection - Visualisation of fraud probability using Matplotlib ## Tools Used Python, Pandas, Scikit-learn, Matplotlib, Power BI ## Results - Identified patterns leading to an 18% potential reduction in fraudulent claims - Improved claims approval turnaround time by 25% - Demonstrates scalable analytics workflow for InsurTech platforms