Detection of Property Fraud in Legal Cases

Detection of Property Fraud in Legal Cases

Author Name : H K Sonam Gowda, Bilvashree H S, Arha A Hegde, Vivek D H, B Uma

DOI: https://doi.org/10.56025/IJARESM.2025.1306251063

ABSTRACT Property fraud sets legal and insurance domains with a tough challenge due to the heavy transaction problems and the subtle nature of fraud patterns. This study aims at the application of the concept of machine learning to detect property-related fraud, considering a real-world scenario supplied by a leading insurance company. We experimented with a variety of supervised learning schemes, including logistic regression, support vector machines, random forest, and XGBoost, in order to pinpoint high-risk cases rapidly and with greater precision. We demonstrate that ensemble methods, with XGBoost in the lead, hold the best prediction power. Explainable AI techniques are used for full transparency and legally interpretable models, so that indicators of fraud such as claim period, renewal behaviour of policies, and inconsistencies in user profiles can be highlighted using SHAP. This framework that is, proposed could dramatically increase early fraud detection, lessen investigative work, and contribute to more dependable legal decision-making.