Research map / Learning theory and algorithms / Clustering
A Better k-means++ Algorithm via Local Search
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 |
|---|---|---|---|
| Add LocalSearch++ rounds (k-means++ sampling with single-swap local search) after k-means++ seeding | k-means++ seeding · k=25 and k=50 on RNA, KDD-BIO, KDD-PHY | clustering cost relative to KM++: 8% to 35% cost reduction | Helped |
| Add 10 Lloyd steps after LocalSearch++ seeding (LL-LS++ vs LL-KM++) | LS++ seeding followed by Lloyd · k=25 and k=50 on RNA, KDD-BIO, KDD-PHY | clustering cost relative to LL-KM++: 1% to 18% reduction, smaller and less stable than without Lloyd | Helped |
| Run LocalSearch++ for many iterations (theory requires Z ≥ 100000 k log log k) | LS++ · theoretical analysis | expected cost approximation | No clear effect |
| Local search time per iteration versus Lloyd iteration | LS++ · empirical running time | runtime | Helped |
Closest papers by meaning
- Fast and Accurate $k$-means++ via Rejection Sampling (NeurIPS 2020 · similarity 0.92)
- Approximate K-Means++ in Sublinear Time (AAAI 2016 · similarity 0.88)
- A Beyond-Worst-Case Analysis of Greedy k-means++ (NeurIPS 2025 · similarity 0.88)
- LSDS++ : Dual Sampling for Accelerated k-means++ (ICML 2023 · similarity 0.88)
- Analyzing $D^\alpha$ seeding for $k$-means (ICML 2024 · similarity 0.86)
- Near-Linear Time Approximation Algorithms for k-means with Outliers (ICML 2024 · similarity 0.83)
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.