Research map / Learning theory and algorithms / Clustering
An Evaluative Measure of Clustering Methods Incorporating Hyperparameter Sensitivity
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Increase the number of hyperparameter evaluations under random search | Clustering algorithms · Generic and NLP datasets | Expected best clustering score (EoM) | Helped |
| Use Bayesian optimization rather than random search to select hyperparameters | Clustering algorithms · Generic datasets | Expected best clustering score (EoM) | Mixed |
| Use Bayesian optimization rather than random search to select hyperparameters | Clustering algorithms · NLP datasets | Expected best clustering score (EoM) | Mixed |
| Use GloVe instead of a BERT-based sentence encoder | Clustering algorithms · NLP datasets | Clustering score | Helped |
| Use DBSCAN instead of HDBSCAN | Density-based clustering · NLP datasets | Clustering score | Helped |
| Use HDBSCAN instead of DBSCAN | Density-based clustering · Generic datasets | Clustering score | Helped |
| Give DBSCAN access to tuned hyperparameters | DBSCAN · Evaluated datasets | Expected best clustering score (EoM) | Helped |
| Fix the number of clusters to the gold count instead of tuning it (ablation) | BIRCH, HAC, and MBKM · AGNews with Glove embeddings | Expected best clustering score (EoM) | Mixed |
| Add BIRCH threshold tuning | BIRCH · NLP datasets | fANOVA importance (%) | Helped |
| Add DBSCAN eps tuning | DBSCAN · NLP datasets | fANOVA importance (%) | Helped |
| Tune HDBSCAN eps, minClusterSize, and minPts together | HDBSCAN · NLP datasets | Clustering score | Helped |
Closest papers by meaning
- Systematic Analysis of Cluster Similarity Indices: How to Validate Validation Measures (ICML 2021 · similarity 0.91)
- A Sampling-Based Approach for Efficient Clustering in Large Datasets (CVPR 2022 · similarity 0.90)
- Efficient Parameter-Free Clustering Using First Neighbor Relations (CVPR 2019 · similarity 0.90)
- Generalization Performance of Ensemble Clustering: From Theory to Algorithm (ICML 2025 · similarity 0.89)
- Ultrametric Cluster Hierarchies: I Want ‘em All! (NeurIPS 2025 · similarity 0.88)
- Average Sensitivity of Hierarchical $k$-Median Clustering (ICML 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.