A Correlation Clustering Approach to Link Classification in Signed Networks -- Full Version --

TL;DR

This paper introduces a novel correlation clustering approach for link classification in signed networks, providing theoretical learning bounds and a new active algorithm with mistake guarantees applicable to arbitrary signed networks.

Contribution

It develops a new theory based on correlation clustering for signed networks and proposes the first mistake-bounded active algorithms for link classification.

Findings

Derived learning bounds for online, batch, and active settings

Introduced a family of efficient active classifiers based on graph circuits

First active algorithms with mistake bounds for arbitrary signed networks

Abstract

Motivated by social balance theory, we develop a theory of link classification in signed networks using the correlation clustering index as measure of label regularity. We derive learning bounds in terms of correlation clustering within three fundamental transductive learning settings: online, batch and active. Our main algorithmic contribution is in the active setting, where we introduce a new family of efficient link classifiers based on covering the input graph with small circuits. These are the first active algorithms for link classification with mistake bounds that hold for arbitrary signed networks.