A Real-Time Framework for Vehicle Insurance Fraud Detection Using Ensemble Learning and Anomaly Detection

View Full Paper

Ask AI

Bookmark

Authors

Discussion

Framework demonstrates high accuracy in detecting vehicle insurance fraud claims, highlighting the role of anomaly detection.

Achieving 92.5% accuracy, the system effectively detects fraudulent vehicle insurance claims in real time.
Using ensemble learning models like XGBoost and LightGBM, the framework minimizes false positives while ensuring high detection rates.
The proposed approach employs anomaly detection via Isolation Forest for identifying irregular claim patterns based on diverse data types.
Evaluation on a dataset of 9,154 claims, where 14.2% were fraudulent, confirms the robustness of the fraud detection methodology.

Prof. Ekta Meshram (2025) studied this question.

View Full Paper

Ask AI

Bookmark