A trust enhanced Recommender System for medicare applications

Recommender Systems are software tools that suggest the users, the required items based on the previous user-item interactions. Becoming extremely powerful, Recommender Systems can be accommodated into areas of vital importance. The proposed work aims to provide healthcare recommendations that include Blood Donor recommendations and Hospital Specialization. The system employs a model based collaborative filtering to predict the future requirements of items, following a specific scenario. The problems in the traditional recommender systems such as cold start and scalability are addressed. The system accommodates a trust factor in the classical recommender system and reaps the efficiencies of the k-means++ algorithm, which provides the threshold rating for the cold start users. The number of clusters required is computed using the slope statistic method. The results of the work show that the proposed system provides cost effective recommendations.