ML-Based Early Detection of Hantavirus Infection Using Clinical and Epidemiological Data

Abstract

The early recognition and diagnosis of Hantavirus infection are key to achieving better clinical outcomes through timely diagnosis and intervention. In this study, an interpretable machine learning framework for patient-level detection of Hantavirus, based on routinely collected clinical and epidemiological data, is proposed. The framework consists of systematic data preprocessing, feature engineering using Mutual Information, and comparative evaluation of 5 supervised machine learning algorithms, namely Logistic Regression, Random Forest, XGBoost, LightGBM, and Support Vector Machine (SVM). To test the robustness of the model, accuracy, precision, recall, F1-score, ROC-AUC, and confusion matrix analysis were used to assess the model's performance. Experimental results showed that the learnt models were consistently highly predictive, and that they were not significantly affected by the varying validation folds, which suggested good model generalisation. The most influential biomarkers associated with Hantavirus infection were determined by the feature importance analysis as follows: CRP, Liver Injury Score, Creatinine, ALT, and AST. The model outlined is a dependable, easily understood strategy for early Hantavirus diagnosis and indicates the requirement for independent clinical testing before its general use in healthcare.