Dissertation/Thesis Abstract

A Local Limit Theorem for Linear Random Fields and Shannon Entropy Estimation for Linear Processes
by Fortune, Timothy W., Ph.D., The University of Mississippi, 2020, 87; 28087868
Abstract (Summary)

We first establish a local limit theorem for a linear random field, which is constructed using a set of real coefficients and a set (referred to as the innovations) of i.id. random variables in L2. Each set is indexed by Zd, where d is the dimension. For each j ∈ Zd, the random variable Xj of the linear random field is constructed as a combination of the innovations using the real coefficients mentioned earlier. We study a sequence Sn of partial sums of the Xj. The manner in which we construct these sums varies and is wholly determined by a sequence of index sets of summation, which we generally refer to as the regions of summation. Under reasonable conditions, we establish the local limit theorem for these sums. That is, we show that the sequence of measures √2π Bn P(Sn ∈ [a, b]) of [a, b] converges to b a, where Bn2 = Var Sn). We also estimate the Shannon entropy S(f) = −E[log f(x)] of a one-sided linear process with probability density function f(x) employing the integral estimator Sn (f), which utilizes the standard kernel density estimator fn(x) of f(x). We show that Sn (f) converges to S(f) almost surely and in L2, under reasonable conditions.

Indexing (document details)
Advisor: Sang, Hailin
Commitee: Dang, Xin, Díaz, Erwin Miña, Bu, Qingying
School: The University of Mississippi
Department: Mathematics
School Location: United States -- Mississippi
Source: DAI-B 82/4(E), Dissertation Abstracts International
Subjects: Statistics
Keywords: Entropy, Field, Limit, Local, Random, Shannon
Publication Number: 28087868
ISBN: 9798678140555
Copyright © 2020 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy