StatREC: a graphical user interface tool for visual hypothesis testing of cost prediction models

  • Authors:
  • Nikolaos Mittas;Ioannis Mamalikidis;Lefteris Angelis

  • Affiliations:
  • Aristotle University, Thessaloniki, Greece;Technological Educational Institute of Kavala;Aristotle University, Thessaloniki, Greece

  • Venue:
  • Proceedings of the 8th International Conference on Predictive Models in Software Engineering
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

Background: During the previous decades there has been noted a significantly increased research interest on the construction of prediction models for accurate estimation of software cost. Despite the development of sophisticated methodologies, there is a continuous debate concerning the divergent and controversial conclusions of the related literature. Nowadays, due to this fact, the research community attempts to systematically base the whole comparison and evaluation process on formal frameworks and structured guidelines in concordance with modern statistical practices and methodologies, so as to resolve the problem of inconsistent findings. Aims: Towards this direction, we present StatREC, a Graphical User Interface, which facilitates the visualization and hypothesis testing of error distributions through their graphical representation as REC curves. Conclusions: The advantage of StatREC is that it provides to the non-expert user a robust, highly interactive, rich in graphics and easily interpretable way to perform comparisons among alternative models. The goal of StatREC is to support project managers during the decision-making process on the cost of software development.