A Team Allocation Technique Ensuring Bug Assignment to Existing and New Developers Using Their Recency and Expertise

Existing techniques allocate a bug fixing team using only previous fixed bug reports. Therefore, these techniques may lead to inactive team member allocation as well as fail to include new developers in the suggested list. A Team Allocation approach for ensuring bug assignment to both Existing and New developers (TAEN) is proposed, which uses expertise and recent activities of developers. TAEN first applies Latent Dirichlet Allocation on previous bug reports to determine the possible bug types. For new developers, TAEN identifies their preferred bug type, and adds them to the list of other developers, grouped under the identified bug types. Upon the arrival of a new bug report, TAEN determines its type and extracts the corresponding group of developers. A heterogeneous network is constructed using previous reports to find the collaborations among the extracted developers. Next, for each developer, a TAEN score is computed combining the expertise and recency of their collaborations. Finally, based on the incoming report’s severity, a team of N members is allocated using the assigned TAEN score and current workloads. A case study conducted on Eclipse Java Development Tools (JDT), shows that TAEN outperforms K-nearest-neighbor Search And heterogeneous Proximity based approach (KSAP) by improving the team allocation recall from 52.88 up to 68.51, and showing the first correct developer on average at position 1.98 in the suggested list. Besides, a lower standard deviation of workloads, 30.05 rather than 46.33 indicates balanced workload distribution by TAEN. Keywords—Bug Assignment; Team Allocation; Bug Report; Latent Dirichlet Allocation (LDA).