Publication Date

2017

Document Type

Dissertation/Thesis

First Advisor

Ebrahimi, Nader B.

Degree Name

Ph.D. (Doctor of Philosophy)

Legacy Department

Department of Mathematical Sciences

LCSH

Statistics; Mathematics

Abstract

In the classical change detection problem, a sequence of independent and identically distributed observations is monitored and at an unknown time the distribution of the data changes in a certain way. The problem is to detect this change as quickly as possible. However, if the assumption of independence is removed the problem is more challenging than the classical change point detection. Especially, if dependence comes from two directions, namely space and time. Moreover, greatly increased availability of data with the presence of dependency over space and time adds further complication and requires thorough investigation. This non-traditional problem assumption provides abundant new research opportunities for finding more efficient detection schemes useful in explaining and understanding the underlying structure of this type of data by accounting for both spatial and temporal dependence among geographic locations. In this dissertation, we provide two competing methods for identifying a single change in dependent over time and space data. Parametric models from spatial data analysis are used to incorporate dependence.

Comments

Advisors: Nader Ebrahimi.||Committee members: Sanjib Basu; Bernard Harris; Duchwan Ryu.||Includes bibliographical references.||Includes illustrations.

Extent

xii, 173 pages

Language

eng

Publisher

Northern Illinois University

Rights Statement

In Copyright

Rights Statement 2

NIU theses are protected by copyright. They may be viewed from Huskie Commons for any purpose, but reproduction or distribution in any format is prohibited without the written permission of the authors.

Media Type

Text

Download

Share

COinS