Consistency of Maximum Likelihood for Continuous-Space Network Models I

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Shalizi, Cosma Rohilla, Asta, Dena Marie
Format: Preprint
Published: 2017
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910772512161792
author Shalizi, Cosma Rohilla
Asta, Dena Marie
author_facet Shalizi, Cosma Rohilla
Asta, Dena Marie
contents A very popular class of models for networks posits that each node is represented by a point in a continuous latent space, and that the probability of an edge between nodes is a decreasing function of the distance between them in this latent space. We study the embedding problem for these models, of recovering the latent positions from the observed graph. Assuming certain natural symmetry and smoothness properties, we establish the uniform convergence of the log-likelihood of latent positions as the number of nodes grows. A consequence is that the maximum likelihood embedding converges on the true positions in a certain information-theoretic sense. Extensions of these results, to recovering distributions in the latent space, and so distributions over arbitrarily large graphs, will be treated in the sequel.
format Preprint
id arxiv_https___arxiv_org_abs_1711_02123
institution arXiv
publishDate 2017
record_format arxiv
spellingShingle Consistency of Maximum Likelihood for Continuous-Space Network Models I
Shalizi, Cosma Rohilla
Asta, Dena Marie
Statistics Theory
Social and Information Networks
Physics and Society
A very popular class of models for networks posits that each node is represented by a point in a continuous latent space, and that the probability of an edge between nodes is a decreasing function of the distance between them in this latent space. We study the embedding problem for these models, of recovering the latent positions from the observed graph. Assuming certain natural symmetry and smoothness properties, we establish the uniform convergence of the log-likelihood of latent positions as the number of nodes grows. A consequence is that the maximum likelihood embedding converges on the true positions in a certain information-theoretic sense. Extensions of these results, to recovering distributions in the latent space, and so distributions over arbitrarily large graphs, will be treated in the sequel.
title Consistency of Maximum Likelihood for Continuous-Space Network Models I
topic Statistics Theory
Social and Information Networks
Physics and Society
url https://arxiv.org/abs/1711.02123