Explainable Graph Spectral Clustering For GloVe-like Text Embeddings

TL;DR

This paper extends explainability methods for graph spectral clustering from simple cosine similarity to more complex embeddings like GloVe, enhancing interpretability of text clustering results.

Contribution

It generalizes previous explainability approaches to include GloVe-like embeddings, broadening the applicability of spectral clustering interpretability.

Findings

Enhanced interpretability of spectral clustering with GloVe embeddings

Demonstrated applicability to various document embedding methods

Improved understanding of clustering results in textual analysis

Abstract

In a previous paper, we proposed an introduction to the explainability of Graph Spectral Clustering results for textual documents, given that document similarity is computed as cosine similarity in term vector space. In this paper, we generalize this idea by considering other embeddings of documents, in particular, based on the GloVe embedding idea.