Predicting defects in SAP Java code: An experience report

Tilman Holschuh, Markus Pauser, Kim Herzig, Thomas Zimmermann, Rahul Premraj, Andreas Zeller
2009 2009 31st International Conference on Software Engineering - Companion Volume  
Which components of a large software system are the most defect-prone? In a study on a large SAP Java system, we evaluated and compared a number of defect predictors, based on code features such as complexity metrics, static error detectors, change frequency, or component imports, thus replicating a number of earlier case studies in an industrial context. We found the overall predictive power to be lower than expected; still, the resulting regression models successfully predicted 50-60% of the
more » ... 0% most defectprone components. * This work was conducted while Tilman Holschuh and Kim Herzig were research interns with SAP, and while Rahul Premraj and Thomas Zimmermann were research associates with Saarland University. Version Archive Defect Database Source Predictor Model Component Quality McCabe FanOut LoC Coupling Predictor Model New Component Component Defect Likelihood Figure 1 . Predicting Defects in a Nutshell. We map previous defects to components and relate the resulting quality to component features. These features can then be used to predict future quality.
doi:10.1109/icse-companion.2009.5070975 dblp:conf/icse/HolschuhPHZPZ09 fatcat:joemoen5xzgcxgdpi527ardysq