Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{197944,
        author = {Dr. MK Jayanthi Kannan and Akanksha Dhote},
        title = {MedSureAI: A Hybrid Ensemble Machine Learning Framework with Anomaly Detection for Real-Time Health Insurance Fraud Detection},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {11},
        pages = {11277-11288},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=197944},
        abstract = {Health insurance fraud has emerged as a critical challenge in modern healthcare systems, resulting in significant financial losses, increased operational costs, and reduced trust among stakeholders. Traditional fraud detection methods, primarily based on rule-based systems and manual auditing, are inadequate for identifying complex and evolving fraud patterns in large-scale healthcare datasets. This research proposes an advanced Artificial Intelligence (AI) driven framework, MedSureAI, designed to enable accurate and real-time fraud detection in health insurance systems. The proposed approach integrates multiple stages, including data preprocessing, advanced feature engineering, hybrid class imbalance handling, and ensemble machine learning techniques. Statistical, behavioral, and risk-based features are extracted to capture hidden fraud patterns. A hybrid imbalance handling strategy combining SMOTE and undersampling is employed to address the skewed distribution of fraudulent and non-fraudulent cases. The model architecture incorporates a combination of Random Forest, XGBoost, and Isolation Forest algorithms, enabling the detection of both known and unknown fraud patterns through supervised and unsupervised learning. The ensemble model demonstrates superior performance compared to individual models, achieving high accuracy, improved recall, and strong ROC-AUC scores. The system is further deployed using a Streamlit-based interface, enabling real-time fraud prediction and automated decision support. Experimental results validate the effectiveness, scalability, and practical applicability of the proposed framework. The study concludes that the integration of hybrid machine learning models with real-time deployment provides a robust and efficient solution for healthcare fraud detection, with potential for further enhancement using advanced AI techniques.},
        keywords = {Health Insurance Fraud Detection, Machine Learning, Artificial Intelligence, Ensemble Learning, Random Forest, XGBoost, Isolation Forest, SMOTE, Class Imbalance Handling, Feature Engineering, Anomaly Detection, Real-Time Prediction, Streamlit Deployment, Healthcare Analytics, Predictive Modeling.},
        month = {April},
        }