2026 Student Thesis : False Claim detection by Machine Learning for Health Insurance, openIMIS

This research presents the design and implementation of an Artificial Intelligence (AI)-based system to detect fraudulent health insurance claims, addressing a critical challenge in modern digital healthcare systems. As insurance platforms like openIMIS expand to support Universal Health Coverage, traditional rule-based claim verification methods are no longer sufficient to identify complex fraud patterns such as overbilling, upcoding, and duplicate claims.

The study develops a data-driven solution using supervised machine learning techniques, specifically Random Forest and XGBoost, to analyze large-scale health insurance data. Using approximately one million anonymized claim records, the system evaluates both administrative and clinical variables, including patient demographics, diagnosis codes, and billing details. To handle the imbalance between legitimate and fraudulent claims, advanced preprocessing techniques such as SMOTE are applied.

A key contribution of the research is the introduction of a probabilistic risk scoring mechanism that replaces binary decision-making with a more flexible and intelligent approach. Claims are categorized into three actionable risk levels: high-risk claims are automatically rejected, medium-risk claims are flagged for manual review, and low-risk claims are automatically approved. This framework significantly improves efficiency by prioritizing human intervention where it is most needed.

In this research we will use Random Forest and XgBoost Machine learning algorithm to compare the result of precission.

The results demonstrate that AI-based models can effectively identify suspicious patterns and enhance the accuracy of claim auditing while reducing administrative workload. The proposed system supports faster claim processing, minimizes financial losses due to fraud, and strengthens the overall sustainability of health insurance programs.

This research provides a practical and scalable framework for integrating AI into health insurance systems, particularly in developing countries, and contributes to the advancement of data-driven governance in healthcare.