Graph Clustering and Community Detection (Unsupervised)

Graph clustering is one of the essential unsupervised tasks. This is a type of graph mining task, and it's helpful in many applications, such as marketing and congestion detection.

K-means clustering

We must first learn the graph embeddings if we use K-means clustering on a graph dataset. Like the previous sections, we can explore DeepWalk to learn the embeddings and then apply K-means clustering.

Let's try to tune the hyperparameters of DeepWalk this time and compute the best performance of graph embeddings on the data of Zachary’s Karate Club.