Drug to Drug Interaction Based on Symptoms

Abstract

In the 21st century, the most important aspect of human life is health care, especially the relation of a doctor and a patient. This relation has a high chance of getting, even more, strengthened up if an AI-based solution is provided for doctors to assist them when dealing with patients in real-time. This project provides an AI-based solution for Doctors. In the current century even though the mortality rate is very low there is still a huge rise in patients’ discomforts and diseases around the world because of which many doctors are pre-occupied and have very little time to give to each patient resulting in a low-quality checkup session for a patient. This project addresses these issues by providing a solution that is based on machine learning and traditional programming techniques. This application uses the drugs dataset provided by Drug-Bank that contains drugs names, interaction details, and adverse effects records. These data set features are used in traditional programming techniques to provide doctors with the drug repository to search through descriptions and suggest medicine based on the patient’s discomforts/disease/indication. This application also uses a machine learning model to predict a patient’s disease based on their current symptoms which further aids in offloading workload from a doctor’s hand resulting in overall better-quality checkup and service for every individual patient. Lastly, this application also eases doctors’ workload by providing a drug interaction checker which uses traditional programming techniques accompanied by a drug database (DrugBank) to evaluate drug interaction to minimize further discomforts for patients.