Inconsistency Detection in Requirement Boilerplates Ontology Using Reasoner and SPARQL

Requirement validation in software requirement engineering is a stage to check whether the specified system requirements are in accordance with customer needs. The consistency check of software requirements is one of the key activities during the requirement validation stage. This activity is generally performed manually by analysts, but it can now be automated using various approaches. However, several issues remain in consistency check activity such as: 1) many inconsistency detection studies remain at the syntactic level; 2) the use of static ontologies that cause limitations in handling system change dynamics; and 3) there has been no research on inconsistency detection of software requirements in requirement boilerplates pattern. Therefore, automation of inconsistency detection in requirement boilerplates dynamic ontology offers a promising solution to facilitate consistency checking. This study solves these problems by building a dynamic ontology-based requirement inconsistency detection software by detecting ontology inconsistencies using the HermiT reasoner and matches ontology components with requirement sentences using SPARQL queries. The test results show a precision of 1.00 and a recall of 0.92, based on reasoning and SPARQL querying over a pre-built dynamic requirement boilerplates ontology. By automating this method through a proof-of-concept software implementation, the process of detecting inconsistencies in requirement boilerplates sentences is expected to become more efficient and accurate.