Research map / Probabilistic and scientific modeling / Variational and generative sampling
Mode-Seeking Clustering and Density Ridge Estimation via Direct Estimation of Density-Derivative-Ratios
Hiroaki Sasaki, Takafumi Kanamori, Aapo Hyvärinen, Gang Niu, Masashi Sugiyama
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 |
|---|---|---|---|
| directly estimate log-density gradient (LSLDGC) instead of KDE-based mean shift with least-squares CV bandwidth | clustering · Sat-image (D=36,n=120,c=6) | ARI: 0.000 -> 0.427 | Improved |
| directly estimate log-density gradient (LSLDGC) instead of mean shift with normal-reference bandwidth | clustering · Sat-image (D=36,n=120,c=6) | ARI: 0.343 -> 0.427 | Improved |
| directly estimate log-density gradient (LSLDGC) instead of mean shift with normal-reference bandwidth | clustering · Speech (D=50,n=400,c=2) | ARI: 0.000 -> 0.146 | Improved |
| LSLDGC instead of MS_NR in low dimension | clustering · Banknote (D=4,n=100,c=2) | ARI: 0.167 -> 0.165 | No clear effect |
| LSLDGC instead of MS_NR | clustering · Olive oil (D=8,n=200,c=9) | ARI: 0.756 -> 0.717 | No clear effect |
| LSLDGC instead of k-means with known cluster count | clustering · Accelerometry (D=5,n=300,c=3) | ARI: 0.499 -> 0.628 | Improved |
| LSLDGC instead of k-means with known cluster count | clustering · Vowel (D=10,n=110,c=11) | ARI: 0.180 -> 0.147 | Hurt |
| LSLDGC instead of spectral clustering | clustering · Speech (D=50,n=400,c=2) | ARI: 0.004 -> 0.146 | Improved |
| coordinate-wise update (LSLDGC_CW) instead of standard fixed-point update | LSLDGC · Banknote (D=4,n=100,c=2) | ARI: 0.165 -> 0.169 | No clear effect |
| direct estimation of ridge (LSDRF) instead of SCMS with least-squares CV bandwidth | density ridge estimation · artificial ridge data | ridge estimation error | Improved |
| LSDRF instead of SCMS with coverage-risk bandwidth | density ridge estimation · artificial ridge data | ridge estimation error | Improved |
Closest papers by meaning
- A Theoretical Analysis of Density Peaks Clustering and the Component-Wise Peak-Finding Algorithm (TPAMI 2023 · similarity 0.89)
- Scalable Laplacian K-modes (NeurIPS 2018 · similarity 0.87)
- TANGO: Clustering with Typicality-Aware Nonlocal Mode-Seeking and Graph-Cut Optimization (ICML 2025 · similarity 0.87)
- Resolution-Aware Structural Density Peak Clustering (NeurIPS 2026 · similarity 0.86)
- Breaking the curse of dimensionality in structured density estimation (NeurIPS 2024 · similarity 0.83)
- Enhanced Denesity Peak Clustering for High-Dimensional Data (AAAI 2025 · similarity 0.82)
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.
