Award Date

12-2010

Degree Type

Thesis

Degree Name

Master of Science in Mathematical Science

Department

Mathematical Sciences

First Committee Member

Chih-Hsiang Ho, Chair

Second Committee Member

Anton Westveld

Third Committee Member

Amei Amei

Graduate Faculty Representative

Billy Bai

Number of Pages

40

Abstract

The task of determining whether a sudden change occurred in the generative parameters of a time series generates application in many areas. In this thesis, we aim at monitoring the change-point of a Poisson process by method, which is characterized by a forward-backward testing algorithm and several overall error control mechanisms. With the application of this proposed method, we declare that Mount Etna is not a simple Poissonian volcano, because two different regimes divided by the change point, January 30th 1974, are identified. The validation procedures, used in a complementary fashion, by the formal hypothesis tests and graphical method will be discussed. In conclusion, the proposed method is easy to implement, and its assessment studies could be conducted based on large scale simulation.

Keywords

Change-point problems; Italy – Mount Etna; Poisson processes; Stochastic processes; Volcanoes

Disciplines

Longitudinal Data Analysis and Time Series | Statistical Models | Statistics and Probability

Language

English


Share

COinS