Dissertation/Thesis Abstract

Statistical analysis of decision trees
by Baskaya, Bengu, M.S., California State University, Long Beach, 2011, 96; 1493089
Abstract (Summary)

Decision trees have been widely used for many years in the statistical literature as powerful, effective, and easily interpretable classification algorithms that are able to automatically select relevant features. This thesis examines in detail both the tree growing phase and the underlying statistical analysis. In addition, various splitting algorithms and stopping rules are explained explicitly to obtain the most powerful tree model. This paper also utilizes numerous techniques to evaluate the accuracy of model. Moreover, the main idea of boosting and its corresponding algorithms (such as AdaBoost, AdaBoost.M1 and AdaBoost.MH) that are used to improve the accuracy of model are also discussed in this paper. Finally, we present two different applications to illustrate how decision tree and boosting method apply to live data.

Indexing (document details)
Advisor: Safer, Alan
Commitee:
School: California State University, Long Beach
School Location: United States -- California
Source: MAI 49/05M, Masters Abstracts International
Source Type: DISSERTATION
Subjects: Statistics
Keywords:
Publication Number: 1493089
ISBN: 9781124622446
Copyright © 2019 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy
ProQuest