Dr. Monica Gracia and Dr. Emmanuel Soriano, researchers at the Universidad Internacional Iberoamericana (UNIB), participate in a study that proposes a new method to reduce the pragmatic ambiguity of natural language in the specifications of operational software development requirements.
Requirements engineering seeks to create products that satisfy the needs of consumers. This process involves activities such as requirements analysis, elicitation, specification, validation and management. However, describing requirements in natural language can lead to ambiguities, which leads to misinterpretation of requirements and can result in a product that does not meet stakeholder expectations. Readers interpret requirements differently depending on their prior knowledge.
When you write requirements specifications in natural language, you are describing the expected functionality of a program prior to its development into operational software. But, at different stages of the process, ambiguities may arise, i.e. situations where several interpretations are possible. These ambiguities can occur at the syntactic level, related to the structure of sentences; semantic, concerning the meaning of words or phrases; domain, linked to the context of the application or specific area; lexical, related to the use of words with multiple meanings; and pragmatic, depending on the context.
Some ambiguity detection approaches exist, however, they do not cover all the necessary concepts and therefore have room for improvement to achieve greater accuracy. Therefore, this study proposes a new approach known as Maximum Concept Matching (MCC) that uses multiple nodes and graph edges of conceptual knowledge. This approach provides a more accurate interpretation of requirements, thus increasing the accuracy in the process of interpretation and detection of pragmatic ambiguity. An algorithm based on edges and nodes has been developed to construct conceptual knowledge graphs and evaluate the effectiveness of this approach in improving the ambiguity detection process in requirements specifications.
The evaluation of the approach showed an accuracy of 65% and recall of 90%, outperforming the existing approach which, on average, has an accuracy of 51% and recall of 63%. These results demonstrate a significant improvement over current methods and support the effectiveness of the proposed approach. In the future, it is intended to improve the performance of this approach by investigating extended paths and an automatic selection of input documents. In addition, it is planned to work with a larger dataset to understand how time and number of requirements influence pragmatic ambiguity detection.
If you want to learn more about this study, click here.
To read more research, consult the UNIB repository.
The Universidad Internacional Iberoamericana (UNIB) promotes several study programs in the area of technology such as the Master in Strategic Management with Specialization in Information Technology. This program focused on business innovation, develops the skills and knowledge of students to manage management positions in systems and ICT, lead organizational change projects using ICT or advise companies seeking to join the competitiveness of new businesses. Do not stay behind, study this program to excel in this field in constant movement and prepare for a bright future!