# A dynamic stochastic blockmodel for interaction lengths

## TL;DR

This paper introduces a novel dynamic stochastic blockmodel that analyzes continuous-time interaction lengths in evolving networks, utilizing a fast variational EM algorithm for inference and demonstrating its effectiveness on real-world face-to-face interaction data.

## Contribution

It presents a new continuous-time dynamic stochastic blockmodel and an efficient inference algorithm, advancing network analysis without time discretization.

## Key findings

- Effective clustering of nodes based on interaction lengths
- Successful application to face-to-face interaction data
- Improved model choice via adapted clustering criterion

## Abstract

We propose a new dynamic stochastic blockmodel that focuses on the analysis of interaction lengths in networks. The model does not rely on a discretization of the time dimension and may be used to analyze networks that evolve continuously over time. The framework relies on a clustering structure on the nodes, whereby two nodes belonging to the same latent group tend to create interactions and non-interactions of similar lengths. We introduce a fast variational expectation-maximization algorithm to perform inference, and adapt a widely used clustering criterion to perform model choice. Finally, we test our methodology on artificial data, and propose a demonstration on a dataset concerning face-to-face interactions between students in a high-school.