Predicting Fraud in Auto Insurance Claims

Predicting Fraud in Auto Insurance Claims

Merugu Bhanu Prasad1, Yellanki Likitha2, Pinninti Sai Varun Reddy3 , Dheeraj S Nair4

1Computer Science & Engineering & Parul University , Vadodara

 2Computer Science & Engineering & Parul University , Vadodara

3Computer Science & Engineering & Parul University , Vadodara

4Computer Science & Engineering & Parul University , Vadodara

Abstract -Fraudulent activities in the domain of auto insurance claims pose significant financial risks to insurance companies, leading to substantial losses and compromised customer trust. This research explores the application of machine learning models, including XGBoost, Logistic Regression, Multi-Layer Perceptron (MLP), and Decision Tree classifiers, to predict fraudulent insurance claims. The study leverages key policyholder attributes, including demographic details, policy specifics, and socio-economic factors, to build robust fraud detection models. The experimental results provide valuable insights into the effectiveness of different machine learning techniques in mitigating fraud risks. Additionally, this paper discusses challenges, regulatory considerations, and real-world applications of AI in fraud prevention.

Key Words: Fraud Detection, Auto Insurance Claims, Machine Learning, XGBoost, Logistic Regression, Multi-Layer Perceptron (MLP), Decision Tree Classifier, Predictive Analytics, Risk Mitigation, Data- Driven Fraud Prevention, Supervised Learning, Anomaly Detection, Regulatory Compliance, Artificial Intelligence (AI) in Insurance, Financial Risk Management.