The Research of Domain Ontology Recommendation Method with Its Applications in Requirement Traceability

Abstract

Ontology can specify the concepts and their relationships. If ontology was used as the bridge to map between the source and target artifacts, it can help to solve the keyword matching problem regarding the requirement traceability in the semantic level. A domain ontology recommendation method has been proposed. Using this method, the lexical semantic representation list of the dependency syntax types can be produced at first using Stanford Parser. Then, the terms as well as their relations can be extracted using the clear algorithm and rule-based matching algorithm. As follows, these terms and their relations can be transformed into domain concepts and their relations, obtaining the domain ontology recommendation. Finally, the domain ontology that is suitable for the requirement traceability can be gained through the processes of selection and revision by the domain experts. The experimental results demonstrate that if the proposed domain ontology was applied to requirement traceability, it can effectively improve the accuracy of the requirement traceability, compared with the vector space model and the requirement traceability method based on HOWNET.