Output details

A Systematic Literature Review on Fault Prediction Performance in Software Engineering

<09> The title is misleading: far from being a straightforward review of the literature, this work is a novel synthesis examining how well different models can perform when predicting defects in code. It follows a similar technique to the Cochran reports, collating data from different published papers and synthesising results that provide new conclusions. Data from this study enabled the more detailed meta-analyses of defect prediction by Shepperd at Brunel, and the article has been referenced by many other studies as providing a benchmark for defect prediction performance. The work also led to a successful EPSRC grant application.