Abstract
Recent research suggests that software testing consumes substantial resources and documenting source code is one way to reduce software maintenance cost. The software engineering literature provides little information on how to establish whether a software component is adequately documented. It appears that benchmarking can be used to identify a software component with little or no documentation. In this paper, we propose and illustrate how a hybrid Data Envelopment Analysis (DEA), Kohonen self-organizing mapping, and Euclidean distance based algorithm can be used to identify poorly documented software components.
Keywords
Get full access to this article
View all access options for this article.