Publication

Abstract : Writing effective requirement specification documents in the face of dynamic user needs is a significant challenge in modern software development. Keyword extraction algorithms can help to retrieve relevant information from functional requirements provided by users. The process of identifying a concise set of words that capture the essence of user stories without losing important information is accomplished through automatic keyword extraction. This paper presents a comparative study of four popular algorithms for keyword extraction: Rapid Automatic Keyword Extraction (RAKE), Yet Another Keyword Extraction (YAKE), TextRank, and KeyBERT. The algorithms are analyzed based on their ability to extract keywords and provide corresponding scores. Additionally, N-gram analysis is performed using the extracted keywords. The study concludes that the RAKE algorithm exhibits better performance in extracting relevant keywords from user stories compared to the other algorithms.