An Explainable AI-Based Decision Support System for Teaching and Classifying Hair Loss Types

This study introduces a leakage-resistant machine learning framework for classifying hair loss types, incorporating explainable artificial intelligence to improve transparency and reliability. The framework uses a unified preprocessing pipeline with nested cross-validation to prevent data leakage and employs SMOTEENN to handle class imbalance. Among several algorithms tested, Extreme Gradient Boosting performed best, achieving an accuracy of 0.8300, F1-score of 0.7908, and AUC of 0.9305 in nested cross-validation, and maintained stable performance on a holdout dataset. The integration of explainable AI allows for interpretable predictions, enhancing the framework's applicability in decision support systems for hair loss classification.