Optimal Nonlinear Prediction of Random Fields on Networks

Abstract

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.

20 pages, 5 figures. For the conference "Discrete Models of Complex Systems" (Lyon, June, 2003). v2: Typos fixed, regenerated figures should now produce readable PDF output

References in corpus (2)

• The structure and function of complex networks
• An Algorithm for Pattern Discovery in Time Series

Cited by in corpus (8)

• Local Causal States and Discrete Coherent Structures
• Quantifying the complexity of random Boolean networks
• Computing With Contextual Numbers
• Algebraic Theory of Patterns as Generalized Symmetries
• A location-mixture autoregressive model for online forecasting of lung tumor motion
• Towards Unsupervised Segmentation of Extreme Weather Events
• Spacetime Symmetries, Invariant Sets, and Additive Subdynamics of Cellular Automata
• Project Dynamics and Emergent Complexity