A Natural Language Processing Technique for Formalization of Systems Requirement Specifications
Viktoria Koscinski, Celeste Gambardella, Estey Gerstner, Mark Zappavigna, Jennifer Cassetti, Mehdi Mirakhorli · 2021
Natural language processing techniques have proven to be useful for analysis of technical specifications documents. One such technique, information extraction (IE), can help automate the analysis of software systems requirement specifications (SysRS) by extracting structured information from unstructured or semi-structured natural language data, allowing for requirements to be converted into formal logic. Current IE techniques are not designed for SysRS data, and often do not extract the information needed for requirements formalization. In this work, we introduce an IE method specifically designed for SysRS data. We provide a description of our approach, analysis of the technique on a set of real requirements, example of how information obtained using our technique can be converted into a formal logic representation, and discussion of our technique and its benefits in automated SysRS analysis tasks.