Trustrace: Mining Software Repositories to Improve the Accuracy of Requirement Traceability Links

Abstract

Abstract-Traceability is the only means to ensure that the source code of a system is consistent with its requirements and that all and only the specified requirements have been implemented by developers. During software maintenance and evolution, requirement traceability links become obsolete because developers do not/cannot devote effort to update them. Yet, recovering these traceability links later is a daunting and costly task for developers. Consequently, the literature proposed methods, techniques, and tools to recover these traceability links semi-automatically or automatically. Among the proposed techniques, the literature showed that information retrieval (IR) techniques can automatically recover traceability links between free-text requirements and source code. However, IR techniques lack accuracy (precision and recall). In this paper, we show that mining software repositories and combining mined results with IR techniques can improve the accuracy (precision and recall) of IR techniques and we propose Trustrace, a trust-based traceability recovery approach. We apply Trustrace on four medium-size open-source systems to compare the accuracy of its traceability links with those recovered using state-of-the-art IR techniques from the literature, based on the Vector Space Model and Jensen-Shannon model. The results of Trustrace are up to 22.7% more precise and have 7.66% better recall values than those of the other techniques, on average. We thus show that mining software repositories and combining the mined data with existing results from IR techniques improves the precision and recall of requirement traceability links.

Autonomous Decentralized Semantic Based Traceability Link Recovery Framework

“…Antoniol, Giuliano et al [21] proposed an approach that discovers traceability links between source code and software engineering artifacts based on Information Retrieval (IR) where software engineering artifacts are ranked against queries constructed from the identifiers of source code. Ali et al [22] proposed an approach to establish and maintain traceability links between source codes and software requirements to improve the precision and recall of information retrieval (IR) techniques by discarding/re-ranking the reported links from IR based on the outcome links that are generated from software repositories. Diaz, Diana, et al [23] proposed an approach that purifies the noise from the candidate links list generated by IR technique.…”

“…Marcus et al [24] proposed approaches that advocate for the use of latent semantic indexing (LSI) to recover traceability links between documentation and source code. However, all the above proposed approaches based on IR, lack accuracy in term of precision, recall, and perform poorly in short context [22], [24], [25]. These approaches also disregard word order, syntactic relations, morphology, semantic relation, and word ambiguities [26].…”

“…If the categories are matched at this level (i.e. current level) then we increment the similarity score by the value of I (line [22][23][24][25][26][27][28][29]. Consider for example that the subject Florida is in artifact 1 and the artifact 2, has the subject Arizona.…”

“…To evaluate AD-STLR, we have chosen one benchmark algorithm from Information retrieval (IR) category while other one from semantic relatedness. Vector Space Model (VSM) [51] was chosen as benchmark because its average precision and recall is significantly better than others algorithms from IR class of traceability link recovery algorithms [7], [9], [10], [12]. Similarly Wu Palmer [20] was chosen for comparison as its precision and recall is close to human judgments and has high performance compared to other semantic similarity algorithms [11], [52], [53].…”

“…During development lifecycle of large scale products when teams are geographically distributed, it becomes impractical to update tractability links continuously. The inability to constantly update the links results in inconsistency in traceability matrix that ultimately results in project overruns, failure, and maintenance difficulty [7]. In the past researchers have focused on traceability link recovery for centralized software development environment by employing mainly Information Retrieval (IR) [8] techniques.…”