Authors: Abhijit Ranjan, Chandraveer Singh
Abstract: In the recent years, the demand for intelligent medication recommender systems has increased tremendously with the evolution of digital health technology. This research targets the development of a symptom-based medication recommender system from a structured and diversified healthcare database. The database is descriptive in nature with information regarding patient symptoms, associated medicines, dietary advice, exercise plans, precautions, and doctor specialties. Early steps of this project include extensive data exploration and preparation from several CSV files to create a clean and solid base for model training. For building the central recommendation engine, traditional machine learning algorithms such as Decision Tree, Random Forest, Naive Bayes, and Logistic Regression were utilized, which try to predict symptoms and suggest the most suitable medicines out of a pre-defined list. Among the models used, the Decision Tree classifier had the best performance, followed by Random Forest, Naive Bayes, and Logistic Regression. The system is smart enough for users to put in symptoms and get suggested medicines for the same, providing useful help for non-emergency medical conditions and employment in resource-constrained environments. Future developments will involve incorporating patient medical history, dosage calculation, drug interaction screening, and implementing the system through a mobile platform to enhance accessibility and real-time use.
DOI: http://doi.org/10.5281/zenodo.17482512
