Use Case Extraction through Knowledge Acquisition

Most challenges in requirements analysis and use case extraction are related to the correct comprehension of clients' core processes and activities, as well as their needs. This information is usually available in documents, such as the business vision, written in natural language. This kind of language may lead to interpretation bias and information loss, thus causing project delays and escalation of costs. In order to overcome natural language interpretation challenges in Requirements Engineering, we propose the use of Verbka, a knowledge acquisition process based on verbal semantics, as a complement to traditional requirements analysis. This process extracts the causal relationships among the actors mentioned in the business vision document. We used this process as a use case pre-modeling tool, aiming to minimize subjectivity in the identification of actors and use cases. We tested Verbka using a business vision from a classical textbook. The results show that this process is able to obtain a list containing all requirements defined by the client, all actors involved in the business vision, and how they interact with each other. This process is systematic and provides a textual and visual representation of user requirements and use cases consistently. Its application reduced interpretation bias, thus allowing a more detailed and structured requirements analysis.