Medicine Recommendation for Pharmacists at Drug Stores based on Data Analysis of Conditional Knowledge

The purpose of this research was to support pharmacists in giving information to patients with medication recommender systems using the Rule-based and Gradient boosting tree methods. The cases of the system consist of the patients taking medication independently without doctors' diagnosis with household medicines to relieve basic symptoms and the symptoms of patients that should receive real treatment at the hospital. In addition, technology solutions such as machine learning have been valuable for application in health care. Accordingly, this research starts with getting a patient's information such as symptoms, chronic diseases, drug allergies, and medicine a pharmacist asks some general questions for understanding the patient's health condition and constructing databases by processing the medication knowledge. Rules were extracted from pharmacy guidelines for filtering the Hospital Cases that are out of the scope of the drugstore and the patient should receive medical care from a specialist. Then the system would analyze the data for recommendations the medicine and display it on the user interface. Lastly, we tested the performance of the system by evaluating with accuracy, precision, recall, and F1 score.