Chapter 9

Entity Resolution and Link Prediction

A network often has missing and duplicated information. For example, one person may be represented by several nodes. One node may hold personal information and another node may hold work information.

Merging these duplicated nodes requires a process called entity resolution. Similarly, there may be missing edges in the graph. The process for finding those edges, called link prediction, can also be used to identify edges that will exist in the future; that is, we can predict which friendships will form between nodes. To do entity resolution and link prediction, there are a number of formulas that rely on the structure of the network and attributes of the nodes. This chapter presents basic methods ...

Get Analyzing the Social Web now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.