An Extensible Recommendation System for Health Research

Abstract

A challenge for basic and clinical biomedical researchers is to determine the particular technologies that can play a role in their research. Core facilities house, centralize, and specialize in particular services and technologies important to biomedical research. At the University of North Carolina at Chapel Hill, over 70 core facilities are available for such use, many with multiple technologies and services. In this paper, we design an Extensible Recommendation System (ERS) to aid investigators in finding which core facilities may impact their research. Recommendation techniques are applied to enhance the process of recognition of technologies and collaborators in a modern biomedical research environment. The architecture of the system has two parts: a core foundation module and extensible module. The foundation module follows from a CMS framework called Plone that handles the basic functions like database queries, user login, etc. The extensible module has a few components such as search component, a recommendation component, etc. The extensible module is designed to be flexible, in that it can be extended by as many recommendation algorithms as we need. In this paper, we compare several recommender systems for biomedical core research facilities. Our simulation results show that the recommender system based on Graph-Theoretic performs best amongst few similar recommenders.