Spectral Clustering Revisited: Information Hidden in the Fiedler Vector
This work offers an incremental improvement for graph clustering applications by enhancing reliability in classification.
The paper tackles the problem of improving spectral clustering by identifying vertices with extreme eigenvector entries as more reliably classified, providing a rigorous proof for the stochastic block model and examples.
We are interested in the clustering problem on graphs: it is known that if there are two underlying clusters, then the signs of the eigenvector corresponding to the second largest eigenvalue of the adjacency matrix can reliably reconstruct the two clusters. We argue that the vertices for which the eigenvector has the largest and the smallest entries, respectively, are unusually strongly connected to their own cluster and more reliably classified than the rest. This can be regarded as a discrete version of the Hot Spots conjecture and should be useful in applications. We give a rigorous proof for the stochastic block model and several examples.