Clinical Decision Support System Using Recurrent Neural Network

This project presents the development of a CDSS (Clinical Decision Support System) utilizing deep learning techniques, specifically focusing on Recurrent Neural Networks (RNNs) and LSTM (Bidirectional Long Short-Term Memory) networks. The system is designed to predict medical conditions based on 132 clinical symptoms and target prognoses. A variety of models, including Support Vector Machine, Decision Tree, K-Nearest Neighbors, deep learning architectures, and Logistic Regression were trained and evaluated for performance using key metrics such as recall, precision, F1-score, and accuracy. Among the models, the Bidirectional LSTM network outperformed traditional machine learning methods demonstrating its more advanced ability to handle sequential clinical data has an accuracy of 90.16%. This research highlights the potential of deep learning models in improving diagnostic accuracy and reliability, offering promising advancements for clinical decision-making systems.