DrugWiser: Machine Learning-Based Personalized Medicine Recommendation System
Robert G. de Luna, Ann Margaret J. Ambasa, Tyrone Paolo V. Garcia, Jowella Marie C. Layao, Jian Louise D. Pelayo, Jhon Mack C. Robledo · 2025
This paper introduces DrugWiser, a recommended system to be implemented using machine learning to provide personalized recommendations of medicine and analyze side effects at the same time. The current study makes use of predictive models such as K-Nearest Neighbors (KNN), Decision Tree Regressor (DTR), Random Forest Regressor (RFR), Support Vector Machine (SVM), Gradient Boosting Regressor (GBR), Ada Boosting Regressor (ABR), and Light Gradient Boosting Machine (LGBM). Random Forest Regressor (RFR) performed the best with its accuracy of 99.98 percent in cross-validation and hold-out validation. To come up with a hybrid recommendation system, the model was combined with content filtering via TF-IDF and Cosine Similarity, and FuzzyWuzzy Matching when the inputs were incomplete. The final prototype was deployed with Raspberry Pi 5 that used Tkinter-based graphical user interface (GUI). Further integration with electronic health records in real time will be a part of future work to develop a better personalized medicine recommendation.