Regularization Logistic Regression Models for Hepatitis B Virus Risk Factors in Algeria, 2025
Aoun Mohamed Bachir, Mezoued FatihaBACKGROUND:
Hepatitis B virus (HBV) remains a major global health challenge, particularly in high-prevalence regions, where early diagnosis and targeted prevention strategies are essential to reduce disease burden and transmission. Strengthening predictive approaches is critical for improving risk-based screening and preventive care.
METHODS:
A cross-sectional study was conducted on 3,000 individuals attending the University Hospital Centre of Beni Messous, Algiers. Advanced regularized regression models, including Least Absolute Shrinkage and Selection Operator, Ridge regression, Elastic Net, and Adaptive Elastic Net, were developed to identify key risk factors and enhance predictive accuracy. Model performance was evaluated using Akaike Information Criterion, Bayesian Information Criterion, cross-validation techniques, and mean squared error.
RESULTS:
The prevalence of HBV infection was 98.3%, with a higher risk observed among older individuals and females. Among the evaluated models, elastic net and Adaptive Elastic Net demonstrated superior predictive performance. The Adaptive Elastic Net model, combined with cross-validation, showed optimal discrimination and stability, enabling accurate identification of high-risk individuals based on demographic and clinical factors.
CONCLUSIONS:
The Adaptive Elastic Net model provides a robust and clinically relevant tool for early detection and risk stratification of HBV infection. From a preventive and strategic medicine perspective, its integration into clinical workflows can enhance targeted screening, support early intervention, and inform data-driven public health strategies aimed at reducing transmission and improving population health outcomes.