Local versus Global Lessons for Defect Prediction and Effort Estimation

Authors:
Tim Menzies;Andrew Butcher;David Cok;Andrian Marcus;Lucas Layman;Forrest Shull;Burak Turhan;Thomas Zimmermann
Affiliations:
West Virginia University, Morgantown;West Virginia University, Morgantown;GrammaTech, Ithaca;Wayne State University, Detroit;Fraunhofer Center, College Park;Fraunhofer Center, College Park;University of Oulu, Oulu;Microsoft Research, Redmond
Venue:
IEEE Transactions on Software Engineering
Year:
2013

Citing 0
Cited 4

The impact of parameter tuning on software effort estimation using learning machines

Proceedings of the 9th International Conference on Predictive Models in Software Engineering
Beyond data mining; towards "idea engineering"

Proceedings of the 9th International Conference on Predictive Models in Software Engineering
A study of subgroup discovery approaches for defect prediction

Information and Software Technology
Leveraging machine learning and information retrieval techniques in software evolution tasks: summary of the first MALIR-SE workshop, at ASE 2013

ACM SIGSOFT Software Engineering Notes

Quantified Score

Hi-index	0.00

Visualization

Abstract

Existing research is unclear on how to generate lessons learned for defect prediction and effort estimation. Should we seek lessons that are global to multiple projects or just local to particular projects? This paper aims to comparatively evaluate local versus global lessons learned for effort estimation and defect prediction. We applied automated clustering tools to effort and defect datasets from the PROMISE repository. Rule learners generated lessons learned from all the data, from local projects, or just from each cluster. The results indicate that the lessons learned after combining small parts of different data sources (i.e., the clusters) were superior to either generalizations formed over all the data or local lessons formed from particular projects. We conclude that when researchers attempt to draw lessons from some historical data source, they should 1) ignore any existing local divisions into multiple sources, 2) cluster across all available data, then 3) restrict the learning of lessons to the clusters from other sources that are nearest to the test data.