Latent Space Models for Dynamic Networks with Weighted Edges

TL;DR

This paper introduces latent space models for dynamic networks with weighted edges, extending binary relational models to count and real-valued data, and demonstrates their effectiveness through simulations and real-world datasets.

Contribution

It develops novel latent space models for weighted dynamic networks using link functions and data augmentation, with inference via MCMC algorithms.

Findings

Models effectively analyze count and real-valued dyads.

Application to mobile phone and trade data reveals network dynamics.

MCMC algorithms successfully estimate model parameters and trajectories.

Abstract

Longitudinal binary relational data can be better understood by implementing a latent space model for dynamic networks. This approach can be broadly extended to many types of weighted edges by using a link function to model the mean of the dyads, or by employing a similar strategy via data augmentation. To demonstrate this, we propose models for count dyads and for non-negative real dyads, analyzing simulated data and also both mobile phone data and world export/import data. The model parameters and latent actors' trajectories, estimated by Markov chain Monte Carlo algorithms, provide insight into the network dynamics.