Local algorithms for interactive clustering

Abstract

We study the design of interactive clustering algorithms for data sets satisfying natural stability assumptions. Our algorithms start with any initial clustering and only make local changes in each step; both are desirable features in many applications. We show that in this constrained setting one can still design provably efficient algorithms that produce accurate clusterings. We also show that our algorithms perform well on real-world data.

References in corpus (2)

• Polynomial Learning of Distribution Families
• Agglomerative Bregman Clustering

Cited by in corpus (8)

• Interactive Bayesian Hierarchical Clustering
• Perceptual Visual Interactive Learning
• Approximate Clustering with Same-Cluster Queries
• The LORACs prior for VAEs: Letting the Trees Speak for the Data
• Interactive Topic Modeling with Anchor Words
• Semi-bandit Optimization in the Dispersed Setting
• Scaling Hierarchical Agglomerative Clustering to Billion-sized Datasets
• Distribution free optimality intervals for clustering