Med-Recommender System for Predictive Analysis of Hospitals and Doctors

Abstract

A recommender system is proposed and developed to help users to find the best hospital for a particular treatment. Finding a best hospital that can cure one’s ailment is of paramount importance. A good hospital is one in which there are always enough staff on duty with the right skills, knowledge and experience. Customer experience is how customers perceive their interactions with a company or an organization. A customer’s experience is reflected in the comments that he makes about the organization through online public forums. Med–recommender system aims to provide accurate analysis of hospitals by taking into account the reviews by thousands of patients, which were written by the patients themselves in various online forums. Our recommendation system performs sentiment analysis on the reviews of various patients using Natural Language Processing techniques to classify them as positive and negative reviews. It weighs the ranking of hospitals on three different parameters namely polarity, subjectivity and intensity. The hospital with the best ranking for curing a particular disease is then given as result to the user asking for a recommendation. The system is evaluated using 300 online reviews about hospitals and specialties and found to yield 90% of accuracy. The proposed system also helps the users to understand the quality of a certain hospital by providing star ratings for the hospital when the user needs.