Labelling issue reports in mobile apps. Issue 6 (1st December 2019)

• Record Type: Journal Article
• Authors: Zhang, Tao; Li, Haoming; Xu, Zhou; Liu, Jian; Huang, Rubing; Shen, Yiran
• Is Part Of: IET software. Volume 13:Issue 6(2019)
• Page Start: 528
• Page End: 542
• Publication Date: 2019-12-01
• Subjects: mobile computing -- learning (artificial intelligence) -- program debugging -- text analysis; feature request -- unlabelled issue report -- bug labelling -- mobile apps -- user reviews -- textual similarity scores -- MCG; Computer software -- Periodicals; Software engineering -- Periodicals; 005.1
• DOI: 10.1049/iet-sen.2018.5420
• Languages: English
• ISSNs: 1751-8806

Abstract

Abstract : Millions of mobile apps have been released to the market. Developers need to maintain these apps so that they can continue to benefit end users, who usually submit issue reports to describe the bugs, the feature requests, and other changes appearing in apps. The labels (e.g. bug, feature request) are important resources to indicate which issue reports should be resolved first or next. According to the investigation, 35.6% of issue reports in top‐17 popular mobile apps are not labelled. Developers have to spend additional time to manually verify each unlabelled issue report so that they can decide to resolve the most important issues. In order to help developers to reduce the workload, in this study, the authors propose a novel approach to automatically tag the unlabelled issue reports. This approach not only computes the similarity between each unlabelled issue report and user reviews related to bugs and features but also calculates the textual similarity scores between each unlabelled issue report and labelled ones. As a result, among all textual similarity measures, this approach using cosine similarity with MCG shows the best performance. Moreover, this approach performs better than the method proposed in the authors' previous study.