Applying Machine Learning and Cloud Computing for Sustainable Healthcare Delivery: Detecting Billing Fraud in the National Health Insurance Scheme in Nigeria

Authors

  • Ngozi Fidelia Dibiaezue

    Department of Software Engineering, University on the Niger, Umunya, Anambra State, Nigeria Department of Computer Science, University of Nigeria, Nsukka, Enugu State, Nigeria
    Author
  • Francis Sunday Bakpo

    Department of Computer Science, University of Nigeria, Nsukka, Enugu State, Nigeria
    Author

DOI:

https://doi.org/10.4314/

Keywords:

Machine Learning; Cloud Computing; Health Insurance; Billing Fraud Detection; Sustainable Healthcare; Anomaly Detection; National Health Insurance Scheme (NHIS)

Abstract

The increase in fraudulent activities in Nigeria’s National Health Insurance Scheme (NHIS) especially fraud by healthcare providers has led to huge financial losses, reduced operational efficiency, and challenges in the provision of sustainable health services. Traditional fraud detection mechanisms in claims management based on manual auditing, static business rules and periodic sampling are slow, labour-intensive and unable to scale with the volume and velocity of electronic claims generated under a digitised health insurance scheme. The paper proposes a machine learning framework hosted on the cloud for automated real time detection of billing fraud in the NHIS claims. The framework combines a supervised ensemble learning model (Random Forest and Gradient Boosting, benchmarked against Logistic Regression and Support Vector Machines) with an anomaly-detection layer to identify both known fraud typologies and previously unseen outlier behaviour, deployed on a scalable cloud-computing infrastructure to enable elastic processing of claims data from distributed healthcare facilities. After pre-processing the historical claims records dataset by cleaning, feature engineering and class-imbalance correction using the Synthetic Minority Oversampling Technique (SMOTE), the models were trained, validated and evaluated using accuracy, precision, recall, F1-score and Area Under the Receiver Operating Characteristic Curve (AUC-ROC). The experimental results demonstrate that the ensemble model deployed on the cloud architecture outperforms the rule-based baseline and individual classical classifiers in terms of fraud-detection performance with significantly higher recall and AUC-ROC. The cloud deployment also reduces the claim-processing latency and improves the horizontal scalability with increasing claim load. The study concludes that the use of machine learning with cloud computing provides a sustainable, cost-effective and scalable pathway to curb billing fraud, thereby protecting the financial integrity of National Health Insurance Schemes and bolstering the broader goal of sustainable healthcare delivery. It recommends policy adoption, data governance and future integration of explainable artificial intelligence (XAI) and federated learning into national claims-management systems.

 

Author Biographies

  • Ngozi Fidelia Dibiaezue, Department of Software Engineering, University on the Niger, Umunya, Anambra State, Nigeria Department of Computer Science, University of Nigeria, Nsukka, Enugu State, Nigeria

     

     

  • Francis Sunday Bakpo, Department of Computer Science, University of Nigeria, Nsukka, Enugu State, Nigeria

     

     

     

Downloads

Published

2026-11-05

How to Cite

Applying Machine Learning and Cloud Computing for Sustainable Healthcare Delivery: Detecting Billing Fraud in the National Health Insurance Scheme in Nigeria. (2026). Communication In Physical Sciences, 13(12), 1994-2016. https://doi.org/10.4314/

Similar Articles

141-150 of 345

You may also start an advanced similarity search for this article.