Recommender System in a Transactional Analytical Solution for Health Care and Health Promotion

Abstract

The Sanologia, born at Havana University, proposes a new conception of people’s health and aims to enrich their lifestyle. The broad development of the information sciences and the maturity that has reached the proposed approach constitute the bases for the creation of new computational solutions. In the present paper, a mathematical-computational solution was conceived and designed for the application of the approach proposed by Sanologia based on NoSQL databases. The solution is characterized to favor the access of the users from diverse locations via mobile devices, from the instrumentation of a distributed architecture and the use of flexible data models. From the operational application, it is feasible to analyze the medical and social data collected in an automated way. Knowledge-based, content-based and collaborative filtering recommendation techniques are hybridized in a recommender system, which using machine learning and natural language processing is incorporated in order to support the decision-making process. The proposed solution is capable of offering personalized recommendations taking into account the users’ historical data, their behavior and the similarity with other users. The validity of the solution was verified by the implementation of a functional prototype and a set of experiments.