Community detection in action

With community detection, our obvious goal is to identify the communities that exist in a network. I explained various approaches in Chapter 9, Community Detection. In this chapter, we will make use of two Karate Club algorithms: SCD and EgoNetSplitter.

For this chapter, and in general, I tend to gravitate toward models that can scale well. If a model or algorithm is only useful on a tiny network, I’ll avoid it. Real networks are large, sparse, and complicated. I don’t think I’ve ever seen something that doesn’t scale well that is actually better than algorithms that do. This is especially true in community detection. The best algorithms do scale well. My least favorite do not scale well at all.

SCD

The first community detection algorithm I want to showcase is SCD. You can find the documentation and journal article about the model at https://karateclub.readthedocs.io/en/latest/modules/root.html#karateclub.community_detection...