High-Performance Classification of Mpox Symptoms Using Support Vector Classifier and Quadratic Discriminant Analysis

Background: Recent global outbreaks of Mpox have posed significant diagnostic challenges, particularly in resource-limited settings. Conventional diagnostic methods are often inaccessible due to cost, logistical constraints, or lack of trained personnel. These limitations highlight the urgent need for alternative, scalable diagnostic strategies. This study explored the application of machine learning (ML) classifiers trained on clinical symptom data as a rapid, cost-effective tool for Mpox detection. Methods: An open-access dataset of clinical symptoms from suspected Mpox cases was used to train and evaluate five supervised ML algorithms: Extra Trees, Quadratic Discriminant Analysis (QDA), Decision Trees, Perceptron, and Support Vector Classifier (SVC). Prior to training, data preprocessing steps, including normalization and handling of missing values, were performed after which model training was carried out using a stratified 80:20 train-test split. Performance was assessed using accuracy, recall, area under the receiver operating characteristic curve (ROC-AUC), and F1-score metrics. Subsequently, feature importance was analyzed using permutation-based techniques to determine the contribution of each clinical symptom to model predictions. Results: Among the five evaluated models, SVC, QDA, and Perceptron achieved superior and identical performance metrics, with accuracy, ROC-AUC, and F1-score values of 97.7%, and a recall of 95.5%. Each of these models correctly identified 44 true positive cases with zero false positives. In addition, QDA and SVC produced the lowest number of false negatives (2) and the highest number of true negatives (42), indicating robust discriminatory power. Feature importance analysis identified skin rash as the most predictive clinical feature, with a permutation importance score of 0.12. Conclusions: These findings demonstrate the strong potential of machine learning classifiers for detecting Mpox based on clinical features. Incorporating these models into healthcare systems could significantly enhance early case detection, improve clinical decision-making, and bolster disease surveillance. Future research should focus on prospective validation of these ML classifiers in real-world clinical environments. Keywords: Mpox, Clinical Symptoms, Machine Learning, Classifiers, Support Vector Classifier, Quadratic Discriminant Analysis, Perceptron, Feature Importance, Clinical Diagnosis.