Multi-Label Classification for Requirement Smell Detection in Natural-Language Requirements
Vilnius University Open Series · 08.05.2026 · 306-316
DOI: 10.15388/lmitt.2026.32 ↗Kurz gefasst
Fachartikel in der Zeitschrift Vilnius University Open Series, online erschienen am 08.05.2026. Verfasst von Karolis Trinkūnas und 1 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Natural-language software requirements often contain quality defects such as ambiguity, vagueness, subjectivity, and nonverifiability. This paper presents a multi-label approach for detecting five requirement smell categories: Subjective, Ambiguous, Nonverifiable, Negative, and Vague. The method adapts SetFit with a SentenceTransformer encoder, a weighted binary-relevance logistic-regression head, hint-aware augmentation, and label-specific threshold tuning. Experiments on harmonized datasets show that the approach supports automated requirement quality analysis and interpretable smell detection.
Zitieren
APA
Trinkūnas, K., & Miliauskaitė, J. (2026). Multi-Label Classification for Requirement Smell Detection in Natural-Language Requirements. Vilnius University Open Series, 306-316. https://doi.org/10.15388/lmitt.2026.32
Vancouver
Trinkūnas K, Miliauskaitė J. Multi-Label Classification for Requirement Smell Detection in Natural-Language Requirements. Vilnius University Open Series. 2026:306-316. doi:10.15388/lmitt.2026.32
BibTeX
@article{Trinknas2026,
author = {Trinkūnas, Karolis and Miliauskaitė, Jolanta},
title = {Multi-Label Classification for Requirement Smell Detection in Natural-Language Requirements},
journal = {Vilnius University Open Series},
year = {2026},
pages = {306--316},
doi = {10.15388/lmitt.2026.32}
}
EUVIMED