PREDICTING THE OBJECT-ORIENTED CLASS STABILITY USING SOFTWARE METRICS

PREDICTING THE OBJECT-ORIENTED CLASS STABILITY USING SOFTWARE METRICS. Masters thesis, King Fahd University of Petroleum and Minerals.

[img]
Preview
PDF
PREDICTING_THE_OBJECT-ORIENTED_CLASS_STABILITY_USING_SOFTWARE_METRICS.pdf

Download (5MB) | Preview

Arabic Abstract

يمثل استقرار الأصناف في البرمجة الموجهة عامل اساسي للمحافظة على جودة البرمجيات. وبالتالي، تم تطوير العديد من المقاييس لقياس درجة استقرار اصناف البرمجة الموجهة. بالرغم من ذلك، فان معظم مقاييس الاستقرار لم تخضع لأي دراسة متعلقة بتوقع استقرار الأصناف. ومن ثم، فان مثل هذه الدراسات تساعد في تحسين جودة البرامجيات، ويمكن استخدامها كمؤشر جودة من أجل المساعدة في الحد منتكاليف صيانة البرمجيات. تسعى هذه الرسالة الى دراسة امكانية استخدام مقاييس الاستقرار كمؤشر للجودة عن طريق توقع استقرار اصناف البرمجة الموجهة.

English Abstract

Class stability in object-oriented systems is an important factor for software quality. Therefore, many class stability metrics have been proposed to measure class stability. Most of the class stability metrics have not been investigated or involved in any prediction studies. Hence, such studies help in improving the software quality and can be used as an early quality indicator of class stability/instability in order to help in reducing maintenance cost and efforts. This thesis seeks to investigate whether the class stability metrics can be used as a predictor for software quality using software metrics. Therefore, prediction models are created to predict the object-oriented class stability. Two artificial intelligence techniques are used, Neural Network and Support Vector Machine, in addition to widely used prediction techniques, Multi Linear Regression and Logistic Regression. Software metrics concerning coupling, cohesion, and complexity are used as predictor variables, where the output of the class stability metrics is used as a target variable of the prediction model.

Item Type: Thesis (Masters)
Subjects: Computer
Department: College of Computing and Mathematics > Information and Computer Science
Committee Advisor: Al-Shayeb, M
Committee Members: Ahmed, M and Niazi, M
Depositing User: Yagoub Eisa (g200903930)
Date Deposited: 10 Jun 2012 07:20
Last Modified: 01 Nov 2019 15:35
URI: http://eprints.kfupm.edu.sa/id/eprint/138671