Automatic use case classification based on topic grouping for requirements engineering.

Requirements Engineering (RE) is the fundamental concept in software engineering scope. Automatic analysis for identifying the requirements can decrease the time consumption, delay, and cost of software development as well. To achieve the best analysis of texts, Natural Language Processing helps to disambiguate the meanings to a great extent. Automatic RE can help much more because automatic RE has more accuracy as well as very low latency in the feature extraction, contrary to manual extraction. The current paper focuses on these issues and suggests a framework for the spontaneous detecting of requirements from customers' documents. The new framework applies to the first stage of requirements analysis. For this reason and according to the evaluation results, the study's new framework could detect topics, use cases, and topic groups by at least 97%, 92%, and 95%, respectively, on the customers' real requirement documents. [ABSTRACT FROM AUTHOR]