AI-BASED DIAGNOSTIC MODEL FOR PEDIATRIC EXANTHEMATOUS DISEASES
Abstract
Differential diagnosis of pediatric exanthematous diseases remains challenging due to overlapping clinical manifestations. To develop and validate an interpretable AI-based diagnostic model for classification of pediatric exanthematous diseases. A retrospective dataset of pediatric patients with confirmed diagnoses (COVID-19, measles, scarlet fever, chickenpox, allergic reactions) was used. A multi-class logistic regression model was developed. Data were divided into training and test subsets (n = 250). Performance was evaluated using accuracy, precision, recall, and F1-score. The overall classification accuracy reached 99.6%. Precision and recall were 100% for most classes and 98% for measles. Validation confirmed stable generalization. The interpretable AI-based model demonstrates high reliability and scalability for integration into clinical decision-support systems.