Model-based Filtering Techniques for Recommendation Systems in Healthcare Domain

A drug recommendation system is a technology-based solution that assists healthcare professionals in suggesting appropriate medications for patients based on various factors such as medical history, symptoms, demographics, and drug effectiveness. The system utilizes advanced algorithms and techniques to analyse large datasets, including patient information, drug profiles, clinical studies, and drug-drug interactions, among others. This research study presents a drug recommendation system that utilizes model-based filtering techniques, specifically Singular Value Decomposition (SVD) and Non-Negative Matrix Factorization (NMF). The system aims to enhance the accuracy and effectiveness of drug recommendations by analysing large diabetes patient datasets containing patient information, drug profiles, and clinical studies. To evaluate the system's performance, three standard evaluation metrics are employed. Furthermore, a visual comparison of the evaluation metrics was presented using a bar graph, which clearly demonstrates the superiority of SVD over NMF. The findings contribute to the advancement of personalized healthcare and emphasize the importance of employing suitable algorithms for accurate medication suggestions.