ABSTRACT
Predicting bug-fixing time plays an important role in allowing a
software manager and team to make decisions about allocation
of resources, prioritization and scheduling. Estimating the time
to fix a bug is not a simple task. In the literature, machine learning
(ML) models have been proposed to help software managers
decide whether a bug might be fixed now or later. One feature
highlighted in ML models for predicting bug-fixing time is reporter
reputation. However, these features are based on the participation
of the reporter or developer in the project, but do not take into
account the time taken to fix the bugs. In this study, we propose
new two features called "reporter rating" and "developer rating."
Unlike reputations, ratings are based on the time taken to fix a
bug. In this study, we carried out an experiment in two datasets
containing bug reports.We ran the reputation and rating features in
ten ML models and compared the results. Additionally, we verified
the features together and combined them with textual features. As
a result, we found that ratings can improve the performance of the
models. Ratings had the best results in probabilistic models, while
reputation was better in models that use the decision tree approach.
When used together, reputations and ratings do not substantially
increase the performance of the models when compared to individual
results. However, ratings improve performance when combined
with textual features more than reputations.