Software Defect Density Analysis

9 pages•Published: September 26, 2019

Abstract

Defect density (DD) is a measure to determine the effectiveness of software processes. DD is defined as the total number of defects divided by the size of the software. Software prediction is an activity of software planning. This study is related to the analysis of attributes of data sets commonly used for building DD prediction models. The data sets of software projects were selected from the International Software Benchmarking Standards Group (ISBSG) Release 2018. The selection criteria were based on attributes such as type of development, development platform, and programming language generation as suggested by the ISBSG. Since a lower size of data set is generated as mentioned criteria are observed, it avoids a good generalization for models. Therefore, in this study, a statistical analysis of data sets was performed with the objective of knowing if they could be pooled instead of using them as separated data sets. Results showed that there was no difference among the DD of new projects nor among the DD of enhancement projects, but there was a difference between the DD of new and enhancement projects. Results suggest that prediction models can separately be constructed for new projects and enhancement projects, but not by pooling new and enhancement ones.

Keyphrases: Defect density analysis, Defect density prediction, ISBSG

In: Frederick C. Harris Jr, Sergiu Dascalu, Sharad Sharma and Rui Wu (editors). Proceedings of 28th International Conference on Software Engineering and Data Engineering, vol 64, pages 139--147

Links:	https://easychair.org/publications/paper/wVQc
	https://doi.org/10.29007/rh9l