Drug Recommender System Based on Cost Minimization Technique for Customized Clinical Prescription

Accessing detailed and comprehensive information about medication, particularly suitability for specific symptoms and severity, remains a daunting challenge for many. Translucency in drug prices by different manufacturers has also been known to cause complications for patients as they struggle to make cheaper choices. Therefore, we propose a sophisticated system to facilitate effective decision-making and cost-effectiveness. The system inputs a person’s symptoms and severity and returns an appropriate medicine’s generic name and possible side effects. The first phase is extracting the appropriate medicine’s generic name from a database through user-inputted symptoms and severity. Afterward, the Extreme Gradient Boosting (XGBoost) machine learning algorithm shall be utilized to provide the system with branded medicine that is cheap enough for the customer. This recommendation considers important factors like prices and the number of users, thereby making data-driven and personalized recommendations, with the system achieving a Mean Squared Error (MSE) of 3.5047. The data set from the Apollo website is very important in the recommendation process. Specifically, it recommends cost-effective branded medicine names based on essential factors like pricing and user popularity. This would emphasize recommending cost-effective branded medicines, ensuring that the economic aspect of healthcare addresses budget constraint-related choices a patient can afford.