In large IS development projects a huge number of unstructured text documents become available and need to be analyzed and transformed into structured requirements. This elicitation process is known to be time-consuming and error-prone when performed manually by a requirements engineer. Thus, previous works have proposed to automate the process through alternative algorithms using different forms of knowledge. While the effectiveness of different algorithms has been intensively researched, limited efforts have been paid to investigate how the algorithms’ outcomes are determined by the utilized knowledge. Our work explores how the amount and type of knowledge affects requirements elicitation quality in two consecutive simulations. The study is based on a requirements elicitation system that has been developed as part of our previous work. We intend to contribute to the body of knowledge by outlining how the provided amount and type of knowledge determine the outcomes of automatic requirements elicitation.
Dieser Eintrag ist Teil der Universitätsbibliographie.
Search Authors in