Chapter 9. Identity Knowledge Graphs

Being able to confidently identify a person or thing is a bedrock of business information systems. Given two records, can you determine if they represent the same real-world thing (a person, an organization, a place, and so on)?

This problem commonly arises when integrating data from different systems, when you need to reason that a record in one system has a counterpart (or not) in the other. In that context, it is often referred to as entity resolution or master data management.

In this chapter, you will learn how knowledge graphs help you address the problem of identity. You will see how the topology of a knowledge graph can give you the confidence to link multiple entities and allow you to reason about them as if they were a single golden record.

Knowing Your Customer

Identity and its myriad of difficulties is part of our everyday experience. For example, Jane Coleman has been a happy and profitable customer of the Bank’s credit card for several years.

Some months ago, Jane married her husband Peter and decided to take his name. Jane Downe and Peter Downe are the owners of a joint checking account with the Bank. But how can the Bank know that Jane Coleman and Jane Downe are the same person? There’s no obligation on Jane to tell the Bank; indeed, she is free to continue to use her maiden name (a not unusual practice among certain professionals like doctors).

As a consequence, Jane regularly gets letters in her mailbox inviting her to open ...

Get Building Knowledge Graphs now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.