Publication Date

2014

Document Type

Dissertation/Thesis

First Advisor

Polansky, Alan M.

Department

Division of Statistics

Abstract

This thesis explores the possibility of testing the statistical significance of modality based on the smoothed bootstrap test for modality. There is little information available on the plausibility of Silverman's test, potentially due to the computational requirements essential to perform the required bootstrap testing. Using the statistical package R, and incorporating parallel processing to perform large bootstrap simulations, the goal is to determine if a statistical test to estimate modality based on significance may exist. Simulation studies reveal properties of critical bandwidth and show modifications to the smoothed bootstrap test for modality make it possible to perform a statistical test for modality.

Extent

79 pages

Language

eng

Media Type

Text

Share

COinS