Interpretable Machine Leaning Models for Healthcare Decision Support

Abstract: Healthcare decision support systems (DSS) play a crucial role in assisting clinicians with diagnosis and treatment planning. However, the lack of interpretability in machine learning models often leads to trust issues, limiting their adoption in clinical settings. This project presents an interpretable machine learningbased decision support system using the Random Forest algorithm to enhance transparency and accuracy in healthcare predictions. The system preprocesses healthcare data, applies feature selection, and generates interpretable insights through feature importance visualization. It integrates with hospital information systems (HIS) for real-time decision-making, ensuring seamless clinical workflow incorporation. Additionally, the project explores natural language processing (NLP) techniques to analyse text-based medical data, improving contextual understanding and decision accuracy. By balancing interpretability and predictive performance, this system enables clinicians to make informed decisions while enhancing trust in AI-driven recommendations. Model validation is conducted using real-world datasets and evaluated through metrics such as accuracy, F1-score, and AUC-ROC. The proposed approach contributes to the development of reliable, interpretable, and efficient healthcare AI solutions that can significantly improve patient outcomes