Graph-Powered Analytics and Machine Learning with TigerGraph

Chapter 7. Better Referrals and Recommendations

This chapter will demonstrate how graph analytics can retrieve information from a network to make better referrals and recommendations, using two real-world use cases. In the first use case, we will build a referral network between patients and healthcare specialists. We will see how to determine which doctors are the most influential and how their interrelations form communities. The second use case is about making a better recommendation engine using features based on the connections and affinities among customers, context factors, products, and features. By the end of this chapter, you should be able to:

• Understand how graph connections provide context

• Apply multiple techniques for analyzing context in order to make recommendations and referrals

• Know how to model and analyze a referral network

• Know how to model and analyze a recommendation engine using graphs

• Explain the meaning of a high PageRank score using the concepts of referral and authority

Case 1: Improving Healthcare Referrals

Today’s healthcare industry has evolved to include many specialties and specialists. This has advanced the state of the art in many areas and given patients the potential to receive expert care. When a patient’s situation is beyond the routine care offered by a general practitioner, the general practitioner may refer the patient to a specialist. There may be subsequent referrals to other specialists. In many healthcare systems, a patient ...