Accession Number : AD1051320

Title :   Causality and Information Dynamics in Networked Systems with Many Agents

Descriptive Note : Technical Report,15 Apr 2016,14 Jan 2017

Corporate Author : University of Waterloo Research Triangle Park United States

Personal Author(s) : Mazumdar, Ravi R

Full Text :

Report Date : 11 May 2017

Pagination or Media Count : 17

Abstract : This report presents results on a theoretical formulation and algorithms for reconstructing Granger causality graphs (GCG) from collections of wide sense stationary (WSS) and cyclostationary time series data. The thrust of the research was to develop methods for GCG sparsification using ideas from Tikhonov regularization and ADMM based proximal algorithms. Several computational examples are presented.

Descriptors :   stochastic processes , random variables , estimators , algorithms , hilbert space , markov processes , graphs , MATHEMATICAL FILTERS , errors , covariance , statistical algorithms , frequency domain , SEQUENCES (MATHEMATICS)

Subject Categories : Statistics and Probability

Distribution Statement : APPROVED FOR PUBLIC RELEASE