Effective Regression Test Case Selection: A Systematic Literature Review

Abstract

Regression test case selection techniques attempt to increase the testing effectiveness based on the measurement capabilities, such as cost, coverage, and fault detection. This systematic literature review presents state-of-the-art research in effective regression test case selection techniques. We examined 47 empirical studies published between 2007 and 2015. The selected studies are categorized according to the selection procedure, empirical study design, and adequacy criteria with respect to their effectiveness measurement capability and methods used to measure the validity of these results. The results showed that mining and learning-based regression test case selection was reported in 39% of the studies, unit level testing was reported in 18% of the studies, and object-oriented environment (Java) was used in 26% of the studies. Structural faults, the most common target, was used in 55% of the studies. Overall, only 39% of the studies conducted followed experimental guidelines and are reproducible. There are 7 different cost measures, 13 different coverage types, and 5 fault-detection metrics reported in these studies. It is also observed that 70% of the studies being analyzed used cost as the effectiveness measure compared to 31% that used fault-detection capability and 16% that used coverage.

DOI: 10.1145/3057269

19 Figures and Tables

Cite this paper

@article{Kazmi2017EffectiveRT, title={Effective Regression Test Case Selection: A Systematic Literature Review}, author={Rafaqut Kazmi and Dayang N. A. Jawawi and Radziah Mohamad and Imran Ghani}, journal={ACM Comput. Surv.}, year={2017}, volume={50}, pages={29:1-29:32} }