An Intelligent Tool for Classifying Issue Reports
2023 (English)In: Proceedings - 2023 IEEE/ACM 2nd International Workshop on Natural Language-Based Software Engineering, NLBSE 2023, Institute of Electrical and Electronics Engineers (IEEE), 2023, p. 13-15Conference paper, Published paper (Refereed)
Abstract [en]
A considerable amount of issue reports are submitted daily in large-scale software development. Manually reviewing and classifying each issue report is challenging and error-prone. Thus, to assist practitioners, in this paper, we propose and evaluate an automatic supervised machine learning-based approach that can automatically predict the newly submitted issue report type (i.e., bug, feature, question, or documentation). We applied the supervised machine learning-based approach to over 1.4 million issue reports data from real open-source projects. We performed our experiments using Stochastic Gradient Descent (SGD)-based classifier and achieved an F1 micro average score of 0.8523. © 2023 IEEE.
Place, publisher, year, edition, pages
Institute of Electrical and Electronics Engineers (IEEE), 2023. p. 13-15
Keywords [en]
Bug Reports, Issue Classification, Natural Language Processing, Software Analytics, Software Maintenance, Gradient methods, Learning algorithms, Learning systems, Natural language processing systems, Open source software, Software design, Supervised learning, Intelligent tools, Language processing, Large-scales, Learning-based approach, Natural languages, Software analytic, Supervised machine learning, Stochastic systems
National Category
Software Engineering
Identifiers
URN: urn:nbn:se:bth-25325DOI: 10.1109/NLBSE59153.2023.00010ISI: 001039169700004Scopus ID: 2-s2.0-85167946589ISBN: 9798350301786 (print)OAI: oai:DiVA.org:bth-25325DiVA, id: diva2:1791346
Conference
2nd IEEE/ACM International Workshop on Natural Language-Based Software Engineering, NLBSE 2023, Melbourne, 20 May 2023
Funder
ELLIIT - The Linköping‐Lund Initiative on IT and Mobile Communications2023-08-252023-08-252024-12-03Bibliographically approved
In thesis