PERSONALIZED HEALTH RECOMMENDATION SYSTEM USING MACHINE LEARNING

Abstract

The collection of health data has become ubiquitous. Wearable devices, fitness applications, and digitized health records culminate into a plethora of health data. While the volume of data continues to grow, the utilization of data in personalizing health recommendations remains scarce. Most recommendations offered are simple, generic, and often misguided. Machine learning can bridge this gap. Unlike advanced rule based systems, modern machine learning systems can dynamically adjust their recommendations as a person’s health status shifts. The focus of this project is the development of a machine learning-based, personalized health recommendation system. It has the ability to integrate disparate data sources to provide meaningful and actionable individual recommendations. These recommendations could stem from fitness wearables, medical history, data on sleep, dietary patterns, and additional health variables. Equally as important is the system’s emphasis on privacy, and trustworthiness, designed to promote sustained reach. Personalization is more flexibility but systems also need to incorporate principles of privacy, trustworthiness, and fairness in order to foster system adoption. The goal of such a system is to transform the healthcare process from reactively where issues are addressed after they arise — to proactive and preventative. With tailored recommendations, people can better understand and detect potential health issues, reducing the need for costly interventions. This is no longer merely an advancement in technology; it is an improvement to an intelligent system of care that puts the users — now more enlightened about their health — at the center of the process.