Building a trust-based doctor recommendation system on top of multilayer graph database.

Abstract

In healthcare applications, developing a data model for storing patient-doctor relationships is important. Though relational models are popular for many commercial and business applications, they may not be appropriate for modeling patient-doctor relationships due to their inherent irregular nature and complexities. In this paper, as a case study, we propose to build a doctor recommendation system for the patients. The recommendation system is built on top of a multilayer graph data model. Contemporary research papers have already shown that multilayer graph data models can be efficiently used in many applications where large, heterogeneous data are to be modeled. As part of the recommendation system, the paper also introduces a concept of trust which is one important ingredient of any kind of recommendation. The trust factor introduced in the paper exploits certain characteristics of the multilayer graph model. The paper also presents some analysis to demonstrate the efficiency of the graph data model in comparison with relational data model.