Hierarchical Similarity Measure for Spectral Clustering
Lucas Siviero Sibemberg, Luiz Emílio Allem, Carlos Hoppen · Proceeding Series of the Brazilian Society of Computational and Applied Mathematics · 2022
In this paper we propose a novel similarity measure for spectral clustering that incorporates a hierarchical component. The main advantage of this measure is that it produces an algorithm that does not depend on any scaling parameter, making it very easy to apply. Our experiments showed that our algorithm performs better than other spectral clustering methods on synthetic data sets with complex shape and multiple scales.