AI Paper Atlas

Research map / Learning theory and algorithms / Clustering

Clustering with Tangles: Algorithmic Framework and Theoretical Guarantees

Solveig Klepper, Christian Elbracht, Diego Fioravanti, Jakob Kneip, Luca Rendsburg, Maximilian Teegen, Ulrike von Luxburg

JMLR 2023 · Learning theory and algorithms › Clustering · approach: Correlation clustering

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 changedWhereMeasuredEffect
raise agreement parameter from 1/3 to 2/3 of smallest cluster size (ablation)Tangles · Binary questionnaire, unbalanced clustersNMIImproved
prune tangle search tree paths of length 1 (ablation)Tangles · Binary questionnaire with useless questionsNMIImproved
add useless random questionsTangles · Binary questionnaire (noise 0.15)NMINo clear effect
increase initial cuts from 20 (2 KL iterations) to 100 (100 iterations) (ablation)Tangles · Synthetic SBM graph clusteringNMI: 20 -> 100No clear effect
increase the number of cuts (ablation)Tangles · Synthetic SBM graph clusteringNMIImproved
use tangles instead of k-means with known k as noise growsk-means · Binary questionnaire, unbalanced clustersNMIMixed
use tangles instead of normalized spectral clusteringSpectral clustering · Synthetic SBM graph clusteringNMINo clear effect
use tangles instead of k-meansk-means · Mixture of 4 GaussiansNMI: 0.797 -> 0.829Improved
use tangles instead of spectral clusteringSpectral clustering · Mixture of 4 GaussiansNMI: 0.805 -> 0.829Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy