Published in Lecture Notes in Computer Science vol. 11622: “Evaluation of Textual Similarity Techniques in Code Level Traceability” (Csuvik, Kicsi and Vidács).
Knowing which tests belong to which code helps quality assurance and maintenance. We examined the textual similarity measures of three machine learning techniques for recovering test-to-code links, with several textual representations of the source code. Evaluated on four open-source systems against naming conventions, the approach improved on the earlier best textual-similarity results for each system.
Read the paper: publisher page
