This web site is no longer maintained and the content may be outdated.
Please visit for up-to-date information.
No upcoming events...

Home / Graduate / M.S. Theses Completed
  Evren Ceylan, 2005    

Thesis Title

Software defect identification using neural network techniques


Software engineering is a tedious job that involves people, tight deadlines and limited budgets. The goal of a software engineer is to develop a software program that produces the desired results of its customers on time and within budget. Delivering what customer wants also involves minimizing the defects in the programs. Defect minimization is the key ingredient in cost management. As the software project progresses, it becomes much more difficult and costly to overcome defect related problems. In such a case there needs to be a long test phase after the development is complete. Depending on the severity of the defects found in the test phase project schedules may be delayed or some fatal defects may end the software development all together. Therefore, it is important to establish quality measures early on in the project life cycle.The main objective of this research is making an analysis in order to overcome these software problems earlier in the project schedule. Such an analysis displays a disciplined process on software development with the aim of making software development more predictable and more efficient. If the software is measured early in its life-cycle, it will be possible to take necessary precautions before undesirable results come out. In order to accomplish this, we propose a model that uses machine learning and artificial neural network methods namely Principal Component Analysis for dimensionality reduction, and Decision Tree, MLP and Radial Basis Functions for defect prediction. The necessary code metric data are obtained from real-life projects of three different software companies in Turkey called Logo Business Solutions, Garanti Teknoloji and Innova Bilişim Sistemleri.
Boğaziçi University Department of Computer Engineering
Address: 34342 Bebek, Istanbul, TURKEY
Phone: +90 212 359 4523-24 Fax: +90 212 287 2461
general information:   webmaster: