Machine Learning-Based Detection of Human Metapneumovirus (HMPV) Using Clinical Data

Human Metapneumovirus (HMPV) is a prominent respiratory pathogen, particularly affecting children, the elderly, and immunocompromised populations. Early detection of HMPV is critical for timely intervention and improved patient outcomes; however, traditional diagnostic methods are often hindered by overlapping symptoms with other respiratory illnesses. This research explores the application of machine learning models for HMPV detection using synthetic clinical data designed to replicate real-world scenarios. The dataset incorporates vital clinical features such as fever, cough, fatigue, symptom duration, oxygen saturation, heart rate, and respiratory rate. To address data imbalance, the Synthetic Minority Oversampling Technique (SMOTE) was employed, resulting in improved sensitivity toward minority class cases. A tuned XGBoost classifier demonstrated robust performance, achieving an accuracy of 73.54%, an F1-score of 0.7063, and a ROC-AUC of 0.7990. Key visualizations, including confusion matrices, ROC curves, and feature importance analyses, provided insights into the model’s efficacy and clinical relevance. This study underscores the potential of machine learning in augmenting clinical decision-making processes for early and accurate detection of HMPV, while also highlighting the importance of preprocessing techniques like data balancing in enhancing model performance. These findings pave the way for scalable, AI-driven diagnostic solutions that can be extended to other respiratory illnesses.