Automating goal-oriented modeling with NLP: extracting actors, roles, and dependencies from functional requirements

Conceptual modeling plays a central role in bridging functional requirements and system design, supporting clear and effective communication among developers, analysts, and stakeholders. This paper proposes an automated approach for generating goal models directly from functional requirements using natural language processing techniques. Goal models capture and structure the high level objectives, intentions, and responsibilities that guide a system or product, thereby strengthening alignment between stakeholder needs and system behavior. By improving clarity, traceability, and decision making during early development phases, goal oriented modeling contributes to systems that better satisfy business objectives and deliver intended value. The proposed method combines NLP preprocessing techniques such as tokenization, part of speech tagging, lemmatization, and stop word removal with rule based extraction of actors, agents, roles, goals, and actor relationships, following the iStar 2.0 framework. The approach is evaluated on more than 500 functional requirements collected from software projects in four different application domains. Experimental results show 94 percent precision, 86 percent recall, and an F1 score of 86 percent for goal model components, demonstrating the accuracy, scalability, and domain independence of the proposed approach.