Defining Classes and Semantic Relationships from User Scenarios Through a Heuristic Approach

End-User Development includes approaches, tools, techniques, and environments that empower end users or non-programmers to create and customize software systems as per their needs without requiring technical expertise. The participation of the end user in transforming the user scenarios into a UML diagram is critical to ensure clarity and completeness of the specification by minimizing ambiguities. The current study proposes a heuristics-based approach to extract the classes and relationships from user scenarios using natural language processing by engaging the end user to ensure the correctness of UML modeling. The End-User Development is integrated with proposed techniques that empower the non-technical users to review, create, and specify the input contextual complexity, enabling seamless software modeling. This study collected 120 user scenarios from five systems of different domains with equal distribution over three categories (i.e., low, medium, and high) to extract the classes and relationships. In the evaluation protocol, the proposed approach, experts, and end users are independent of each other for all user scenarios. The results based on the accuracy measure revealed that the proposed approach performs 22% and 13% better than the evaluation performed by the end users and experts, respectively, highlighting the need for the proposed approach to act as a supportive tool to enhance further engagement in EUD.