Improving Wart Treatment Efficacy Prediction with Immunotherapy and Cryotherapy Using SVM, RF, CVM, k-NN, MLP, and BLR Algorithms in Machine Learning

Warts, characterized by small, rough, benign tumors induced by human papillomavirus (HPV), pose a challenge in treatment prediction due to variations in patient response and disease characteristics. While a machine learning-based expert prediction system and associated rules have been proposed, their success remains unsatisfactory and lacks interpretability for medical professionals. In this study, we employ a decision tree-based approach to delineate rules for predicting wart treatment success. Our findings reveal success rates ranging from 90% to 95% across different treatment methods, surpassing previous reports. Additionally, the decision tree rules generated can be converted into visual representations, enabling intuitive interpretation of treatment success rates in relation to patient age and disease duration. This method offers a straightforward and more precise interpretation of rules for medical experts, facilitating the prediction of treatment success as a percentage.