Integrating Machine Learning for Predictive Epidemiological Modeling of Toxoplasma Gondii Infection in Occupationally Exposed Populations of Lahore
Journal of Contemporary Insights in Health Sciences · 2025
DOI: 10.69591/jcihs.3.1.1 ↗Abstract
Background: Toxoplasma gondii infection poses a significant health risk, particularly among occupationally exposed populations. Early detection and risk prediction are crucial for effective public health interventions. Objective: This study aims to develop a machine learning-driven approach to predict and evaluate the risk of Toxoplasma gondii infection among occupationally exposed employees in Lahore, Pakistan. Methods: A total of 120 participants, including 60 sewage workers, 30 gardeners, and 30 construction workers, were assessed using biological assays (ELISA and PCR) alongside socio-demographic information, including age, education, hygiene practices, and pet ownership. Three supervised learning algorithms, Logistic Regression, Decision Tree, and Random Forest were applied to model infection risk. Model performance was evaluated using accuracy, precision, recall, F1-score, and ROC-AUC. Analysis of feature importance identified the key predictors of infection Results: The Random Forest classifier outperformed other models, achieving 92% accuracy and an AUC of 0.90. The analysis revealed that cat ownership cat ownership, poor hygiene, and low educational level were the strongest predictors of infection risk. Conclusion: Integrating machine learning with traditional serological assays provides a reliable, data-driven framework for early detection and risk stratification of Toxoplasma gondii infection. This approach can inform targeted public health interventions for high-risk occupational groups.
EUVIMED