An Effective Cardiac Care Recommender System

Abstract

Recommender systems are used now-a-days as information filtering tools for predicting user’s preferences, predominantly from big data and recommends things to their taste/choice. Recommender System for health domain provides a mean to assist users in categorizing users with similar symptoms of diseases. Users often feel uncertain when investigating on their own about the medical information that is available on the various online sources. The primary objective is to suggest the preferred cardiologist to the users based on the symptom parameters they have given as input. This information is analyzed and users with similar medical information are identified. Different users may have different expectations in the selection of doctors. So users are asked to prioritize their expectations. The proposed system is based on both collaborative and content filtering approach that makes use of the information provided by users, analyzes them and identifies the similarity between users through the clusters created using K-Means algorithm. In K-Means clustering, the two types of clusters will be formed based on the health parameters given by the users and the other one will be formed based on the users priority. The similar users that are available in both the clusters are identified and ratings of those similar users are taken into matrix. The doctor with the maximum average is recommended. Based on the ratings for the doctors given by the similar users and priority of their expectations, doctors are recommended. We strongly believe that the idea of recommending the doctors based on the users similarity and priority can be adapted to cope with the special requirements of the health domain.