Rule-based system for automated classification of non-functional requirements from requirement specifications

Unmasking the non-functional requirements (NFRs) such as quality attributes, interface requirements and design constraints of software is crucial in finding the architectural alternatives for software starting from early design opinions. For developing quality software product, extraction of NFRs from requirement documents is needed to be carried out and it's beneficiary if this process becomes automated, reducing the human efforts, time and mental fatigue involved in identifying specific requirements from a large number of requirements in a document. The proposal presented in this paper combines automated identification and classification of requirement sentences into NFR sub-classes with the help of rule-based classification technique using thematic roles and identifying the priority of extracted NFR sentences within the document according to their occurrence in multiple NFR classes. F1-measure of 97% is obtained on PROMISE corpus and 94% F1-Measure on Concordia RE corpus. The results established validates the claim that proposal provides specific and higher results than the previous state of art approaches.