The automotive industry is known for its stringent quality and safety requirements, especially in the development of complex software systems. Software Requirements (SRs) are the foundational artifacts that define the behavior, architecture and functionality of these systems, making their quality crucial to the success of any project. However, studies have shown that project failures are often not due to technical issues, but to poorly defined or misunderstood SRs. This challenge is exacerbated in the automotive industry, where increasing software complexity and shorter development timelines require more efficient processes to ensure SR quality. To address these issues, this master thesis presents the development of Hofer Requirements Analysis (HORA), a solution that automates the detection of quality degrading elements in SRs. By leveraging Natural Language Processing (NLP) and Machine Learning (ML) techniques, the HORA tool identifies “requirements smells” such as ambiguity, vagueness and subjectivity that may compromise SR clarity and consistency. NLP techniques such as tokenization, Part of Speech (POS) tagging and syntactic analysis are used to automatically flag problematic terms in SRs, thus reducing the need for extensive manual review. The HORA tool was evaluated against annotated ground truth dataset. Its performance was measured against key metrics such as accuracy, precision, recall and F1 score. HORA demonstrated its effectiveness in improving the quality of SRs by achieving 86% accuracy, 50% precision, 83% recall and 62% F1 score. In this way, HORA not only increases the efficiency of SR inspections, but also contributes to the development of higher quality, more secure software systems in the automotive industry.
| Datum der Bewilligung | 2024 |
|---|
| Originalsprache | Englisch |
|---|
| Betreuer/-in | Thomas Schlechter (Betreuer*in) |
|---|
- Automotive Mechatronics and Management
Quality Analysis Tool for Software Requirements by using AI Techniques
Dora, R. D. (Autor). 2024
Studienabschlussarbeit: Masterarbeit