Virtual Medical Assistant Using Machine Learning

Abstract

In the current scenario, people are very much hesitant to visit hospitals in fear that they might catch an infection. This fear escalates when they are made to wait at the hospital for a long time. Also, doctors find it difficult to write or type prescriptions with gloves and protective clothing on. The aim of this project is to design and develop an application which acts as a virtual assistant to the doctor. The application at the patient’s end would ask patients to register themselves for an appointment. They are asked to record their symptoms. The application then extracts the symptom related words from the recording and predicts the possible diseases. The extracted symptoms and diseases assist the doctor in taking decisions. The doctor could listen to the recording if he/she wants to. The doctor could ask the patient to take some tests, if the patient needs any, and also schedule a call with the patient if necessary. The system would also help the doctor record his prescription by voice, and convert it to prescription format and send it to the patient. This application would save the doctor’s time and energy and avoid misinterpretation of drug names, thereby simplifying the whole process. The proposed system will be developed in python, applying principles of Natural Language Processing, Keyword Extraction and Machine Learning.