Hierarchical Models for Relational Event Sequences

TL;DR

This paper introduces a hierarchical modeling approach for analyzing multiple relational event sequences, enabling better inference of event dynamics and variation across groups, demonstrated through prediction experiments and a classroom dynamics case study.

Contribution

It presents a novel hierarchical extension for relational event sequence models, allowing shared information across sequences and improved inference.

Findings

Hierarchical models improve prediction accuracy.

Effective sharing of information across sequences.

Application to high school classroom dynamics.

Abstract

Interaction within small groups can often be represented as a sequence of events, where each event involves a sender and a recipient. Recent methods for modeling network data in continuous time model the rate at which individuals interact conditioned on the previous history of events as well as actor covariates. We present a hierarchical extension for modeling multiple such sequences, facilitating inferences about event-level dynamics and their variation across sequences. The hierarchical approach allows one to share information across sequences in a principled manner---we illustrate the efficacy of such sharing through a set of prediction experiments. After discussing methods for adequacy checking and model selection for this class of models, the method is illustrated with an analysis of high school classroom dynamics.