Consistencies and inconsistencies between model selection and link prediction in networks

Toni Vallès-Català, Tiago P. Peixoto, Marta Sales-Pardo, Roger Guimerà

Research output: Contribution to journalArticlepeer-review

41 Citations (SciVal)
100 Downloads (Pure)

Abstract

A principled approach to understand network structures is to formulate generative models. Given a collection of models, however, an outstanding key task is to determine which one provides a more accurate description of the network at hand, discounting statistical fluctuations. This problem can be approached using two principled criteria that at first may seem equivalent: selecting the most plausible model in terms of its posterior probability; or selecting the model with the highest predictive performance in terms of identifying missing links. Here we show that while these two approaches yield consistent results in most cases, there are also notable instances where they do not, that is, where the most plausible model is not the most predictive. We show that in the latter case the improvement of predictive performance can in fact lead to overfitting both in artificial and empirical settings. Furthermore, we show that, in general, the predictive performance is higher when we average over collections of models that are individually less plausible than when we consider only the single most plausible model.

Original languageEnglish
Article number062316
Pages (from-to)1 - 12
Number of pages12
JournalPhysical Review E
Volume97
Issue number6
DOIs
Publication statusPublished - 28 Jun 2018

Funding

This work was supported by a James S. McDonnell Foundation Research Award (M.S.-P. and R.G.), and by Grants No. FIS2016-78904-C3-1-P (M.S.-P. and R.G.) and No. FIS2015-71563-ERC (to R.G.) from the Spanish Ministerio de Economía y Comptetitividad. APPENDIX A:

ASJC Scopus subject areas

  • Statistical and Nonlinear Physics
  • Statistics and Probability
  • Condensed Matter Physics

Fingerprint

Dive into the research topics of 'Consistencies and inconsistencies between model selection and link prediction in networks'. Together they form a unique fingerprint.

Cite this