In this chapter, we studied different ways of defining and measuring node importance, also known as centrality, either using the number of connections for each node (degree centrality, PageRank, and its derivatives) or path-related metrics (closeness and betweenness centrality).
In order to use these algorithms from GDS, we also studied different ways to define the projected graph, which is the graph that's used by GDS to run the algorithm. We learned how to create this projected graph using both native and Cypher projection.
In the last section of this chapter, we saw how centrality algorithms can help in the practical application of fraud detection, assuming fraudsters are more likely to interact with each other.
A related topic is the concept of communities or patterns within a graph. We will investigate this in the next chapter. We'll use different types of algorithms to find communities or clusters within a graph in an unsupervised or semi-supervised way and identify...
