SpEx: A Spectral Approach to Explainable Clustering

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Argov, Tal, Wagner, Tal
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866918182023856128
author Argov, Tal
Wagner, Tal
author_facet Argov, Tal
Wagner, Tal
contents Explainable clustering by axis-aligned decision trees was introduced by Moshkovitz et al. (2020) and has gained considerable interest. Prior work has focused on minimizing the price of explainability for specific clustering objectives, lacking a general method to fit an explanation tree to any given clustering, without restrictions. In this work, we propose a new and generic approach to explainable clustering, based on spectral graph partitioning. With it, we design an explainable clustering algorithm that can fit an explanation tree to any given non-explainable clustering, or directly to the dataset itself. Moreover, we show that prior algorithms can also be interpreted as graph partitioning, through a generalized framework due to Trevisan (2013) wherein cuts are optimized in two graphs simultaneously. Our experiments show the favorable performance of our method compared to baselines on a range of datasets.
format Preprint
id arxiv_https___arxiv_org_abs_2511_00885
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle SpEx: A Spectral Approach to Explainable Clustering
Argov, Tal
Wagner, Tal
Machine Learning
Data Structures and Algorithms
Explainable clustering by axis-aligned decision trees was introduced by Moshkovitz et al. (2020) and has gained considerable interest. Prior work has focused on minimizing the price of explainability for specific clustering objectives, lacking a general method to fit an explanation tree to any given clustering, without restrictions. In this work, we propose a new and generic approach to explainable clustering, based on spectral graph partitioning. With it, we design an explainable clustering algorithm that can fit an explanation tree to any given non-explainable clustering, or directly to the dataset itself. Moreover, we show that prior algorithms can also be interpreted as graph partitioning, through a generalized framework due to Trevisan (2013) wherein cuts are optimized in two graphs simultaneously. Our experiments show the favorable performance of our method compared to baselines on a range of datasets.
title SpEx: A Spectral Approach to Explainable Clustering
topic Machine Learning
Data Structures and Algorithms
url https://arxiv.org/abs/2511.00885