Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


cooper final.pdf (769.71 KB)

ETD Abstract Container

Abstract Header

Change Point Analysis for Lognormal Distribution Based on Schwarz Information Criterion

Cooper, Richard
http://orcid.org/0000-0003-3676-2296
http://rave.ohiolink.edu/etdc/view?acc_num=bgsu1594660837226868

Abstract Details

2020, Master of Arts (MA), Bowling Green State University, Mathematics.
Data is categorical or numerical information that is collected and analyzed. Some examples of data are the company’s financial data or environmental data that can inform us of the state of the environment over time. Circumstances often lead to change. For example, the temperature globally has risen dramatically over the years, which began a phenomenon referred to as global warming. As a result, glaciers are rapidly melting, causing sea levels to rise. Rising sea levels damaging coastal areas with increase flooding and more frequent hurricanes. Nevertheless, if we can discover when the change occurred and what caused it, we can put reg-ulations and procedures in place to help avoid future damages. Determining the number of changes in a data set and when it occurred, is called the change point problem. Akaike (1973) introduced the Akaike information criterion (AIC) to solve the change point problem. Additional information criterion was introduced to expand upon and improve AIC. We will discuss AIC further, and we will utilize the Schwarz Information Criterion (SIC) and the Modified information criterion (MIC) to perform simulations on random lognormal distributed data. Chapter 1 introduces AIC and the binary segmentation procedure. Binary segmentation detects multiple change points in a data set. Note, the hypothesis test is employed to perform both AIC and binary segmentation procedures. Chapter 2 introduces SIC, then a theoretical basis and mathematical formulas are derived. Next, we use the criterion to administrate multiple simulations such as power tests and multiple sample size comparisons. Then we utilize SIC and the binary segmentation procedure to analyze the annual net flux of carbon to the atmosphere from land-use. Chapter 3 introduces MIC. Again, theoretical basis and mathematical formulas are derived. Multiple simulations are conducted, such as power test, sample size comparison, and information approach comparison. Lastly, we analyze historical cryptocurrency financial data using MIC with the binary segmentation procedure. This thesis aims to study the performance of the model selection information criterion and analyze the power of the change point detection procedure. We will do that through multiple simulations, conducted in chapters 2 and 3.
Wei Ning, Dr. (Advisor)
John Chen, Dr. (Committee Member)
73 p.
Statistics
Change Point Analysis; Akaike information criterion; Schwarz Information Criterion; Modified information criterion

Recommended Citations

Citations

  • Cooper, R. (2020). Change Point Analysis for Lognormal Distribution Based on Schwarz Information Criterion [Master's thesis, Bowling Green State University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=bgsu1594660837226868

    APA Style (7th edition)

  • Cooper, Richard. Change Point Analysis for Lognormal Distribution Based on Schwarz Information Criterion. 2020. Bowling Green State University, Master's thesis. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=bgsu1594660837226868.

    MLA Style (8th edition)

  • Cooper, Richard. "Change Point Analysis for Lognormal Distribution Based on Schwarz Information Criterion." Master's thesis, Bowling Green State University, 2020. http://rave.ohiolink.edu/etdc/view?acc_num=bgsu1594660837226868

    Chicago Manual of Style (17th edition)

bgsu1594660837226868
319
© 2020, all rights reserved.
This open access ETD is published by Bowling Green State University and OhioLINK.