Medical Recommendations: Leveraging CRNN with Self-Attention Mechanism for Enhanced Systems

As a result of the exponential increase in the global population, many individuals have limited access to medical care. Robust medicine recommendation systems are needed to provide timely and accurate drug recommendations, improving patient outcomes and optimizing healthcare utilization. This research paper aims to propose a recommendation system for medical drugs that utilize sentiment analysis and captures main features using CRNN with a self-attention mechanism. This self-attention was instrumental in enabling the system to concentrate on critical segments of the text, while the CRNN architecture facilitated the capture of both spatial and temporal features of the input. The system yielded 96% accuracy, alongside precision, recall, and F1 scores of 95.4%, 93.8%, and 94.6%, surpassing multiple baseline models and hybrid models in recommending top drugs accurately for particular conditions. Hence, by providing efficient and effective drug recommendations, the proposed system can significantly benefit medical professionals and patients, particularly those in underserved communities.