Using sensitivity analysis to create simplified economic models for regression testing

  • Do H
  • Rothermel G
  • 29

    Readers

    Mendeley users who have this article in their library.
  • 27

    Citations

    Citations of this article.

Abstract

Software engineering methodologies are subject to complex cost-benefit tradeoffs. Economic models can help practitioners and researchers assess methodologies relative to these tradeoffs. Effective economic models, however, can be established only through an iterative process of refinement involving analytical and empirical methods. Sensitivity analysis provides one such method. By identifying the factors that are most important to models, sensitivity analysis can help simplify those models; it can also identify factors that must be measured with care, leading to guidelines for better test strategy definition and application. In prior work we presented the first comprehensive economic model for the regression testing process, that captures both cost and benefit factors relevant to that process while supporting evaluation of these processes across entire system lifetimes. In this work we use sensitivity analysis to examine our model analytically and assess the factors that are most important to the model. Based on the results of that analysis, we propose two new models of increasing simplicity. We assess these models empirically on data obtained by using regression testing techniques on several non-trivial software systems. Our results show that one of the simplified models assesses the relationships between techniques in the same way as the full model.

Author-supplied keywords

  • economic models
  • empirical studies
  • regression test selection
  • regression testing
  • sensitivity analysis
  • test case prioritization

Get free article suggestions today

Mendeley saves you time finding and organizing research

Sign up here
Already have an account ?Sign in

Find this document

Get full text

Authors

  • Hyunsook Do

  • Gregg Rothermel

Cite this document

Choose a citation style from the tabs below

Save time finding and organizing research with Mendeley

Sign up for free