Clustering-based Automated Requirement Trace Retrieval

Abstract

The benefits of requirement traceability are well known and documented. The traceability links between requirements and code are fundamental in supporting different activities in the software development process, including change management and software maintenance. These links can be obtained using manual or automatic means. Manual trace retrieval is a time-consuming task. Automatic trace retrieval can be performed via various tools such as Information retrieval or machine learning techniques. Meanwhile, a big concern associated with automated trace retrieval is the low precision problem primarily caused by the term mismatches across documents to be traced. This study proposes an approach that addresses the term mismatch problem to obtain the greatest improvements in the trace retrieval accuracy. The proposed approach uses clustering in the automated trace retrieval process and performs an experimental evaluation against previous benchmarks. The results show that the proposed approach improves the trace retrieval precision.

An Enhancing Recovery Links between Two Artifacts Using Variational Autoencoder

“…Second: comparing the proposed trace retrieval approach results with three studies (i.e., studies [6] and [47], and [47]). Study ( [6]) utilized unsupervised machine learning-based clustering, Study ( [47]) utilized active learning, and study ( [48]) utilized supervised machine learning. First: Results of the proposed trace retrieval approach based VAE.…”

“…This section reviews some methods used to solve the term mismatch problem in automated trace retrieval between requirements and source code. In [6], the authors proposed an approach that addresses the term mismatch problem between requirements and source code to obtain the most significant improvements in trace retrieval accuracy. The proposed approach used unsupervised machine learning based on the clustering in the automated trace retrieval process and performed an experimental evaluation against previous benchmarks.…”

“…The authors previously acknowledged certain problems associated with three datasets (eTour, SMOS, and eANCI) in their research (See ref. [6]). Specifically, all three datasets are imbalanced, and to address this issue, the authors proposed oversampling algorithm for deep learning models called SMOTE, which based on the widely used SMOTE method.…”

“…However, manual retrieval of traceability links can be error-prone [4] and timeconsuming. Therefore, automatic retrieval techniques that utilize tools such as information retrieval [5], ontology, machine learning [6], and deep learning [7] are often employed. Deep learning approaches can be classified into two main categories: supervised learning and unsupervised learning [8].…”

“…This highlights the need to focus on improving precision. Consequently, several researchers have explored various approaches to enhance precision [6,10]. However, these attempts have only yielded marginal improvements.…”