Hands-On Machine Learning with C++

Chinese Whispers – graph clustering algorithm with Dlib

The Chinese Whispers algorithm is an algorithm to partition the nodes of weighted, undirected graphs. It was described in the paper Chinese Whispers—an Efficient Graph Clustering Algorithm and its Application to Natural Language Processing Problems by Chris Biemann. This algorithm also does not use any unique graph theory methods but it uses the idea of using local contexts for clustering, so it can be classified as a density-based method.

In the Dlib library, this algorithm is implemented in the chinese_whispers() function, which takes the vector of weighted graph edges and outputs the container with cluster indices for each of the vertices. For the performance consideration, we limit ...