Reviewing rounds prediction for code patches

Abstract Code review is one of the common activities to guarantee reliability software, while code time-consuming as it requires reviewers inspect source each patch. A patch may be reviewed more than once before eventually merged or abandoned, and then such a tighten development schedule developers further affect progress project. Thus, tool that predicts early on how long will can help take self-inspection beforehand for patches require long-time review. In this paper, we propose novel method, PMCost , predict reviewing rounds uses number features, including meta-features, diff personal experience features textual better reflect changes process. To examine benefits perform experiments three large open projects, namely Eclipse, OpenDaylight OpenStack. The encouraging experimental results demonstrate feasibility effectiveness our approach. Besides, study why proposed contribute prediction.