Affordable Access

Publisher Website

Joining Ranked Inputs in Practice**The support of the National Science Foundation under Grants IIS-0093116, EIA-9972883, IIS-9974255, and EIA-9983249 is gratefully acknowledged.-Chapter 87

DOI: 10.1016/b978-155860869-6/50094-9
  • Biology
  • Computer Science
  • Logic
  • Mathematics
  • Medicine


Publisher Summary This chapter carries out an extensive performance study to evaluate two recent algorithms for obtaining a global rank from multiple ranked inputs. A growing number of database applications require the processing of ranking queries based on multiple attributes. In the context of multimedia retrieval, predicates often involve image similarity matching with respect to several features. Users may present an example image, and query the database for images “most similar” to the example based on color and texture. Although each database image object can easily be ranked for color and texture separately, results must be presented to the user in a combined similarity order. Another example from information retrieval is the search for documents containing search topics from multiple sources. While each source provides retrieved documents sorted by relevance, the collection of retrieved documents returned to the user must be sorted in a combined relevance order. Other examples include information retrieval based on multiple keywords, CAD similarity searches, geometrical databases, medical imaging, and molecular biology.

There are no comments yet on this publication. Be the first to share your thoughts.