Home // IMMM 2021, The Eleventh International Conference on Advances in Information Mining and Management // View article

Predicting Software Quality from Development and Release Factors

Authors:
Rishita Mullapudi
Tajmilur Rahman
Joshua Nwokeji

Keywords: Software Quality, Release Quality, Software Quality Model, Open Source Software, Decision Trees

Abstract:
Long lasting sustainable systems require quality software releases. If a new version of the software encounters relatively fewer post-release defects, i.e., bugs, then we can consider that version as a better quality release. In the competitive world of faster release and shorter release cycle based development, it is challenging to deliver a quality release of a software product. Predicting the release quality certainly helps developers to take precautions and measures to prevent post-release bugs. Although many researchers studied software quality prediction, a lack of robust empirical study on software development historical data to predict their impact on software release quality has been observed. In this study, we predict the release quality of Eclipse Equinox project by constructing a decision tree model from six factors, such as code changes (churns), commits, churns in test-files, churns in config-files, last-minute-change, etc., observed from the historical data extracted from the version control system. Such development and release factors will give us a better understanding on how the developers’ activities affect the quality of a software release. Five quality levels, i.e., classes are used in our classification model from the Eclipse bugs depending on the presence of different levels of severity of bugs. Furthermore, we will construct three more models, Naıve Bayes, K-means Clustering, and Linear Regression, and will compare the accuracy of prediction. The outcome of this study will be a set of classification models built on the six development factors and an insightful comparison among them.

Pages: 1 to 6

Copyright: Copyright (c) IARIA, 2021

Publication date: May 30, 2021

Published in: conference

ISSN: 2326-9332

ISBN: 978-1-61208-864-8

Location: Valencia, Spain

Dates: from May 30, 2021 to June 3, 2021