Cosma Rohilla Shalizi - Optimal Nonlinear Prediction of Random Fields on Networks

dmtcs:2310 - Discrete Mathematics & Theoretical Computer Science, January 1, 2003, DMTCS Proceedings vol. AB, Discrete Models for Complex Systems (DMCS'03) - https://doi.org/10.46298/dmtcs.2310
Optimal Nonlinear Prediction of Random Fields on NetworksArticle

Authors: Cosma Rohilla Shalizi 1

  • 1 Center for the study of complex systems

It is increasingly common to encounter time-varying random fields on networks (metabolic networks, sensor arrays, distributed computing, etc.).This paper considers the problem of optimal, nonlinear prediction of these fields, showing from an information-theoretic perspective that it is formally identical to the problem of finding minimal local sufficient statistics.I derive general properties of these statistics, show that they can be composed into global predictors, and explore their recursive estimation properties.For the special case of discrete-valued fields, I describe a convergent algorithm to identify the local predictors from empirical data, with minimal prior information about the field, and no distributional assumptions.


Volume: DMTCS Proceedings vol. AB, Discrete Models for Complex Systems (DMCS'03)
Section: Proceedings
Published on: January 1, 2003
Imported on: November 21, 2016
Keywords: recursive estimation,Networks,random fields,sufficient statistics,nonlinear prediction,information theory,[INFO.INFO-DM] Computer Science [cs]/Discrete Mathematics [cs.DM],[MATH.MATH-CO] Mathematics [math]/Combinatorics [math.CO],[NLIN.NLIN-CG] Nonlinear Sciences [physics]/Cellular Automata and Lattice Gases [nlin.CG],[INFO.INFO-HC] Computer Science [cs]/Human-Computer Interaction [cs.HC]

10 Documents citing this article

Consultation statistics

This page has been seen 276 times.
This article's PDF has been downloaded 272 times.