Home // ICSEA 2015, The Tenth International Conference on Software Engineering Advances // View article

Towards a Better Understanding of Static Code Attributes for Defect Prediction

Authors:
Muhammed Maruf Öztürk
Ahmet Zengin

Keywords: Defect prediction; Low level metrics; Metric derivation

Abstract:
Defect prediction requires intensive effort and includes operations which are focused on reducing the cost of software development. These operations involving the use of machine learning algorithms could produce wrong results originated from skewed or missing data. In order to increase the success rate of predictors, defect data sets are either pruned or duplicated. To address this problem, we observe the effects of the derivation of low level metrics using statistical methods in prediction performance. The performance of predictions are evaluated using 10-fold cross-validation on each data set. Experimental results obtained by using 15 data sets show that naive Bayes classifier improved values of Area Under the Curve (AUC) with the rate of 0,1 in average.

Pages: 40 to 44

Copyright: Copyright (c) IARIA, 2015

Publication date: November 15, 2015

Published in: conference

ISSN: 2308-4235

ISBN: 978-1-61208-438-1

Location: Barcelona, Spain

Dates: from November 15, 2015 to November 20, 2015