Combining textual and structural analysis of software artifacts for traceability link recovery

Author

McMillan, Collin ; Poshyvanyk, Denys ; Revelle, Meghan

Author_Institution

Dept. of Comput. Sci., Coll. of William & Mary, Williamsburg, VA

fYear

2009

fDate

18-18 May 2009

Firstpage

41

Lastpage

48

Abstract

Existing methods for recovering traceability links among software documentation artifacts analyze textual similarities among these artifacts. It may be the case, however, that related documentation elements share little terminology or phrasing. This paper presents a technique for indirectly recovering these traceability links in requirements documentation by combining textual with structural information as we conjecture that related requirements share related source code elements. A preliminary case study indicates that our combined approach improves the precision and recall of recovering relevant links among documents as compared to stand-alone methods based solely on analyzing textual similarities.

Keywords

information retrieval; software engineering; system documentation; requirements documentation; software documentation artifacts; source code elements; structural analysis; textual analysis; traceability link recovery; Computer science; Documentation; Educational institutions; Indexing; Information analysis; Information retrieval; Large scale integration; Matrix decomposition; Software tools; Terminology;

fLanguage

English

Publisher

ieee

Conference_Titel

Traceability in Emerging Forms of Software Engineering, 2009. TEFSE '09. ICSE Workshop on

Conference_Location

Vancouver, BC

Print_ISBN

978-1-4244-3741-2

Type

conf

DOI

10.1109/TEFSE.2009.5069582

Filename

5069582