Inferring social structure from continuous-time interaction data

TL;DR

This paper introduces a novel approach to infer latent social networks from continuous-time relational event data by focusing on deviations from expected interaction patterns, rather than just interaction frequency.

Contribution

It proposes an alternative to point process models by emphasizing persistent social connections and deviations, enhancing understanding of underlying social structures.

Findings

Effective identification of latent social ties in college student data.

Successful inference of social structure in barn swallow interaction data.

Demonstrated advantages over traditional contagion-based models.

Abstract

Relational event data, which consist of events involving pairs of actors over time, are now commonly available at the finest of temporal resolutions. Existing continuous-time methods for modeling such data are based on point processes and directly model interaction "contagion," whereby one interaction increases the propensity of future interactions among actors, often as dictated by some latent variable structure. In this article, we present an alternative approach to using temporal-relational point process models for continuous-time event data. We characterize interactions between a pair of actors as either spurious or that resulting from an underlying, persistent connection in a latent social network. We argue that consistent deviations from expected behavior, rather than solely high frequency counts, are crucial for identifying well-established underlying social relationships. This study aims to explore these latent network structures in two contexts: one comprising of college students and another involving barn swallows.