Special Issue Article
An Efficient Approach for Requirement Traceability Integrated With Software Repository
|Related article at Pubmed, Scholar Google|
Traceability links between requirements of a system and its source code are helpful in reducing system conception effort. During software updates and maintenance, the traceability links become invalid since the developers may modify or remove some features of the source code. Hence, to acquire trustable links from a system source code, a supervised link tracing approach is proposed here. In proposed approach, IR techniques are applied on source code and requirements document to generate baseline traceability links. Concurrently, software repositories are also mined to generate validating traceability links i.e. Histrace links which are then called as experts. Now a trust model named as DynWing is used to rank the different types of experts. DynWing dynamically assigns weights to different types of experts in ranking process. The top ranked experts are then fed to the trust model named as Trumo. Trumo validates the baseline links with top ranked experts and finds the trustable links from baseline links set. While validating the links, Trumo is capable of discarding or re-ranking the experts and finds most traceable links. The proposed approach is able to improve the precision and recall values of the traceability links.