Research map / Learning theory and algorithms / Clustering
Kernel k-Groups via Hartigan’s Method
Guilherme Franca, Maria L. Rizzo, Joshua T. Vogelstein
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| replace Lloyd's heuristic (kernel k-means) with Hartigan's method (kernel k-groups) | kernel clustering · Dermatology | accuracy: 0.872 -> 0.962 | Improved |
| use kernel k-groups instead of kernel k-means | kernel clustering · UCI datasets | NMI: 0.867 -> 0.928 | Improved |
| use kernel k-groups instead of kernel k-means | kernel clustering · UCI vehicle | NMI: 0.166 -> 0.126 | Hurt |
| use kernel k-groups instead of kernel k-means | kernel clustering · UCI forest | NMI: 0.457 -> 0.450 | Hurt |
| refine Bethe Hessian output with kernel k-groups | Bethe Hessian · Girvan-Newman SBM (lambda=1.5) | overlap: 0.840 -> 0.870 | Improved |
| refine Bethe Hessian with kernel k-groups at undetectable lambda=0.6 | Bethe Hessian · Girvan-Newman SBM (lambda=0.6) | overlap: 0.183 -> 0.177 | No clear effect |
| use kernel k-groups instead of Bethe Hessian as number of communities k grows | Bethe Hessian · sparse SBM (n=1000, d=3) | overlap | Improved |
| use kernel k-groups instead of Bethe Hessian | Bethe Hessian · arXiv GR-QC | modularity: 0.46 -> 0.55 | Improved |
| use random or k-means++ initialization instead of Bethe Hessian initialization (ablation) | kernel k-groups · sparse SBM | overlap | Hurt |
| use kernel k-groups instead of GMM for highly unbalanced clusters | kernel k-groups · unbalanced Gaussian mixture | accuracy | Hurt |
| use kernel k-groups instead of k-means and GMM on non-Gaussian 1D data | kernel k-groups · 1D two-class non-Gaussian | accuracy: 0.516 -> 0.851 | Improved |
| use kernel k-groups instead of kernel k-means with many clusters | kernel k-groups · 2D grid clusters | accuracy | Improved |
Closest papers by meaning
- Multiple Kernel Clustering with Kernel k-Means Coupled Graph Tensor Learning (AAAI 2021 · similarity 0.87)
- Graph-based Clustering Revisited: A Relaxation of Kernel k-Means Perspective (JMLR 2026 · similarity 0.87)
- Scalable Multiple Kernel Clustering: Learning Clustering Structure from Expectation (ICML 2024 · similarity 0.86)
- Analysis of spectral clustering algorithms for community detection: the general bipartite setting (JMLR 2019 · similarity 0.86)
- Dynamic Similarity Graph Construction with Kernel Density Estimation (ICML 2025 · similarity 0.85)
- COKE: Core Kernel for More Efficient Approximation of Kernel Weights in Multiple Kernel Clustering (ICML 2025 · similarity 0.84)
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.
