Software-Change Prediction: Estimated+Actual

Huzefa Kagdi, Jonathan Maletic
2006 2006 Second International IEEE Workshop on Software Evolvability (SE'06)  
The authors advocate that combining the estimated change sets computed from impact analysis techniques with the actual change sets that can be recovered from version histories will result in improved software-change prediction. An overview of both impact analysis (IA) and mining software repositories (MSR) is given. These are compared and a discussion of their expressiveness and effectiveness is presented. A framework is proposed to integrate these two approaches for software-change prediction.
doi:10.1109/software-evolvability.2006.14 fatcat:s3bgkrybcfcdvkzkrq355gd7ma