Predicting the Next Change at the Fine-Grained Level


Changing source code is not an easy task. Developers occasionally change source code incorrectly or overlook code fragments that need to be changed. Such mistakes entail additional cost in having to reedit the source code correctly, and repeated changes themselves can be a hazard to software quality. We are conducting research into realizing automated code changing as a countermeasure for human errors. As the first step of this research, in this paper we propose a technique to predict what kinds of program elements will be deleted and added in the next change to Java methods. We also evaluate two types of prediction using the proposed technique. One is predicting all types of program elements. The other is predicting only program elements that are statements or larger elements. The precision of the two types of prediction were experimentally found to be 54%-96% and 73%-99%, respectively.

12 Figures and Tables

Cite this paper

@inproceedings{Murakami2014PredictingTN, title={Predicting the Next Change at the Fine-Grained Level}, author={Hiroaki Murakami and Keisuke Hotta and Yoshiki Higo and Shinji Kusumoto}, year={2014} }