Statistical data analysis for software metrics validation

  • Authors:

  • Ming-Chang Lee

  • Affiliations:

  • Department of Information Management, Fooyin University, Taiwan

  • Venue:
  • KES'05 Proceedings of the 9th international conference on Knowledge-Based Intelligent Information and Engineering Systems - Volume Part IV
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

A metrics validation process is defined that integrates quality factors and quality functions. It proposes a comprehensive metric validation methodology that has validity criteria, which support the quality function and activities conducted by software organization for the purpose of achieving project quality goals. In this paper, valid metrics are assessing differences in quality, assessing relative quality, control quality (discrimination between high quality and low quality), control quality (tracking changes), and prediction quality. The criteria are defined and illustrated by association, consistency, discriminative power, tracking. Statistical methods such as Mann-Whitency, Wilcoxon Rank Sum test, Wald-Wolfowitz, and Discriminate Analysis play an important role in evaluating metrics against the validity criterion.