Recommending Developers For Issues In a Social Cording Platform

Abstract

Numerous open source platforms exist which generally accommodate an open bug repository. Extensive open source projects are usually strained by the rate of incoming bug reports. Any new issue documented must be analysed to verify if it is relevant, and assigned to a particular developer for it to be resolved. However, any manual approach to do so takes up time and resources. To address this issue, similarity checking was performed between the technologies familiar to developers and the technologies required to resolve an issue on GitHub. Furthermore, developer activeness and quality were determined through engagement metrics and contribution attributes to assist the generation of a ranked list of recommended developers for an issue. The system was evaluated by domain experts in addition to the accuracy, performance and scalability testing performed.