Author Type

Graduate Student

Date of Award

1994

Document Type

Thesis

Publication Status

Version of Record

Submission Date

October 2026

Thesis/Dissertation Advisor [Chair]

Khoshgoftaar, Taghi M.

Abstract

One goal of software engineers is to produce software products. An additional goal, that the software production must lead to profit, releases the power of the software product market. This market demands high quality products and tight cycles in the delivery of new and enhanced products. These market conditions motivate the search for engineering methods that help software producers ship products quicker, at lower cost, and with fewer defects. The control of software defects is key to meeting these market conditions. Thus, many software engineering tasks are concerned with software defects. This study considers two sources of variation in the distribution of software defects: software complexity and enhancement activity. Multivariate techniques treat defect activity, software complexity, and enhancement activity as related multivariate concepts. Applied techniques include principal components analysis, canonical correlation analysis, discriminant analysis, and multiple regression analysis. The objective of this study is to improve our understanding of software complexity and software enhancement activity as sources of variation in defect activity, and to apply this understanding to produce predictive and discriminant models useful during testing and maintenance tasks. These models serve to support critical software engineering decisions.

Extent

167 p.

Comments

College of Engineering and Computer Science|FAU Electronic Theses and Dissertations Collection|Thesis (Ph.D.)--Florida Atlantic University, 1994.|College of Engineering and Computer Science

Identifier

FADT12383

Format

pdf

Download

Share

COinS