8 Constructing a nearest neighbor similarity network

 

This chapter covers

  • Manually extracting node features
  • Presenting network motifs and graphlets
  • Introducing betweenness and closeness centralities
  • Constructing a monopartite network based on pairwise cosine similarities
  • Using the community detection algorithm to complete a user segmentation task

This chapter will describe constructing a similarity network based on node properties or features. Like a typical machine learning preprocessing workflow, each data point or node is represented as a vector. In the machine learning context, a vector is a list of one or more numerical values. When dealing with graphs, there are generally two approaches you could take to describe a node as a vector. You could manually produce a set of features that describes a node, or you could use various graph algorithms to produce vectors representing a node in the network. In this chapter, you will manually create representations of nodes to describe their roles in the network and then use those representations to construct an inferred similarity network.

8.1 Feature extraction

8.1.1 Motifs and graphlets

8.1.2 Betweenness centrality

8.1.3 Closeness centrality

8.2 Constructing the nearest neighbor graph

8.2.1 Evaluating features

8.2.2 Inferring the similarity network

8.3 User segmentation with the community detection algorithm

8.4 Solutions to exercises

Summary

sitemap