Date of Award

8-2009

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Legacy Department

Mathematical Science

Committee Chair/Advisor

Gallagher, Colin M

Committee Member

Kulasekera , Karunarathna B

Committee Member

Sun , Xiaoqian

Abstract

The problem of undocumented change-points in data sets appears in many areas of science. Mathematical fundamentals of asymptotic methods used in change-point analysis are discussed, and several important maximally selected change-point statistics are introduced. First, the likelihood ratio method is applied to abstract data models within the setting of precipitation series. Basic inference as to the legitimacy and effectiveness of asymptotic methods at detecting undocumented change-points is provided. Next, maximally selected chi-square statistics are discussed in detail and applied to data on tropical cyclone behavior, where a widely available and widely analyzed data set on Atlantic basin cyclones is studied. Change-points in several indicators are found, and importantly a recent increase in cyclone frequency is verified. Next, several methods for detecting undocumented mean shifts in correlated data are introduced and analyzed. Here, we determine that LR-type tests are preferable to CUSUM tests. Finally, propositions regarding change-points in Markov chains and the full likelihood ratio test for ARMA process are discussed.

Share

COinS
 
 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.