AI Paper Atlas

Research map / Graph learning / Graph algorithms and clustering

Clustering from General Pairwise Observations with Applications to Time-varying Graphs

Shiau Hong Lim, Yudong Chen, Huan Xu

JMLR 2017 · Graph learning › Graph algorithms and clustering · approach: Neural and local graph 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
use MLE weights instead of linear or uniform weights (ablation)convex clustering (Algorithm 1) · Sparse planted partition graphs (n=200, n=1000)full recovery rateImproved
use linear weights instead of MLE weights (ablation)convex clustering (Algorithm 1) · Dense planted partition graphsfull recovery rateMixed
use negative (conservative) deviation of weights from the MLE mean (ablation)convex clustering with inexact weights · Gaussian labels (n=200, 8 clusters, mu=2 and 1.5)full recovery rateImproved
add random noise in [-Delta, Delta] to the MLE weights (ablation)convex clustering · Gaussian labels (n=200, 8 clusters)full recovery rateHurt
increase cluster size K with total nodes fixed (ablation)convex clustering · Planted partition (n=400)full recovery rateImproved
make observations highly certain instead of 0.7-reliable (ablation)convex clustering · Partial observations, n=600 to 2400, K=150fraction of observed entries for 90% recovery: 0.7 reliable: linear in n; certain: sub-linear in nImproved
add more snapshots T (ablation)convex clustering on time-varying graphs · Independent and Markov snapshots (n=200, 8 clusters)full recovery rateImproved
replace evolutionary spectral clustering with the convex programtime-varying graph clustering (T=1,2,3) · Synthetic time-varying graphs (fully observed)pairwise error rate: higher computational costImproved
use Markov snapshot model instead of independent or union models (ablation)convex clustering · Reality Mining (n=85, 14 weekly snapshots)pair accuracyImproved
replace Markov convex method with evolutionary spectral clusteringReality Mining clustering · Reality Mining (n=85)pair accuracyImproved

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