Improving Reliability of Defect Prediction Models

Developing error free software is a prime objective of software developers. However, achieving this objective is not a trivial task. This becomes even more harder for the development organizations with minimum resources. The developers often use automatic defect prediction models, which are mostly developed using machine learning algorithms, to locate defects in software. The prediction quality of such models is important as wrong predictions may negatively impact on the development organizations as well as the end users. This book explores the possible issues in the existing prediction models and proposes methods to further enhance the prediction quality of such models.

Autorentext

Jayalath Ekanayake arbeitet derzeit als Dozent für Informatik an der Uva Wellassa University, Sri Lanka. Sein Hauptforschungsinteresse gilt dem Mining von Software-Repositories.