An experimental evaluation of SEC+, an enhanced search engine for component-based software development

  • Authors:
  • Sofien Khemakhem;Khalil Drira;Emna Khemakhem;Mohamed Jmaiel

  • Affiliations:
  • LAAS-CNRS, Toulouse Cedex;LAAS-CNRS, Toulouse Cedex;University of Sfax, Sfax, Tunisia;University of Sfax, Sfax, Tunisia

  • Venue:
  • ACM SIGSOFT Software Engineering Notes
  • Year:
  • 2008

Quantified Score

Hi-index 0.00

Visualization

Abstract

Current approaches for service discovery are inherently restricted to the exact querying. This may provide incomplete answers since queries are often overspecified and may lead to low precision and recall. To alleviate these problems, we achieved an experimental evaluation that uses of the enhanced search engine, SEC+. This engine is based on the subsumption mechanism and a function that calculates the semantic distance. Both the used rate and the non-functional features are considered to filter the selection. We show that such a solution can improve the quality of the search and can enhance both the recall and the precision.