CLUBench: A Clustering Benchmark
Year: 2026
Type: conference-paper
Abstract: Clustering is a fundamental problem in data science with a long-standing research history, yielding numerous insightful algorithms. Despite this progress, a systematic and large-scale empirical evaluation that jointly considers conventional algorithms, deep learning-based methods and recent foundation model-based clustering remains largely absent, leading to limited guidance on algorithm selection and deployment. To address this gap, we introduce CLUBench, a comprehensive clustering benchmark comprising 24 algorithms of diverse principles evaluated on 131 datasets across tabular, text, and image data, involving 178,815 experiments. Importantly, our analysis on (i) the impact of hyperparameter tuning, (ii) the impact of data types and characteristics, (iii) the impact of pretrained embeddings, (iv) large language model-based clustering, (v) the similarity of algorithms, and (vi) low-rank structures of performance matrices, yields meaningful insights and promising pathways about clustering research. For instance, our study reveals that: 1) All evaluated deep clustering methods do not exhibit a significant advantage compared with the top-performing conventional clustering algorithms (e.g., KMeans, SpeClu) in terms of average performance on tabular datasets; 2) For image and text clustering tasks, combining pretrained embeddings with conventional clustering algorithms (e.g., KMeans, SpeClu) offers an effective and efficient clustering; 3) Clustering remains a challenging and nontrivial problem, even in the era of increasingly dominant foundation models. Moreover, we propose to use the low-rank structure in cross-model performance matrices to efficiently approximate the overall performance evaluation in practical applications. We further demonstrate the feasibility of model selection based on the performance matrices across all hyperparameter configurations. Finally, we provide an easy-to-use toolbox by encapsulating the source code from the official code repository into a unified framework, accompanied by detailed instructions. All benchmark datasets, overall experimental results, and the toolbox are fully open-sourced and available at https://github.com/xiaofeng-github/CLUBench.
Authors: Feng Xiao, Dazhi Fu, Chris Ding, Jicong Fan
Institution: Chinese University of Hong Kong, Shenzhen
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7196968120