Automatic sentence annotation for more useful bug report summarization
University of Lethbridge. Faculty of Arts and Science
Lethbridge, Alta. : University of Lethbridge, Department of Mathematics and Computer Science
Bug reports are a useful software artifact with software developers referring to them for various information needs. As bug reports can become long, users of bug reports may need to spend a lot of time reading them. Previous studies developed summarizers and the quality of summaries was determined based on human-created gold-standard summaries. We believe creating such summaries for evaluating summarizers is not a good practice. First, we have observed a high level of disagreement between the annotated summaries. Second, the number of annotators involved is lower than the established minimum for the creation of a stable annotated summary. Finally, the traditional fixed threshold of 25% of the bug report word count does not adequately serve the different information needs. Consequently, we developed an automatic sentence annotation method to identify content in bug report comments which allows bug report users to customize a view for their task-dependent information needs.
bug reports , automatic sentence annotation , annotated summaries , bug report summarization , sentence labelling , Software failures -- Documentation -- Abstracts , Natural language processing (Computer science) , Automatic abstracting , Text processing (Computer science) , Abstracts -- Data processing , Dissertations, Academic