Ranking

In some kinds of information retrieval systems may either descriptors or search term be given different weights.

 

The search result consequently is not a simple question of either-or but a question of degrees. Displays or prints can thus be ranked according to the weights of the terms in the records. It is also possible to define "cut-off-values". Alternatively may prints per default contain, for example, the 100 highest ranked records/documents.

 

A similar technique is used in Internet search engines. Google's "PageRank" algorithm is a well known example, which is link-based (see also Meghabghab, 2001 and Thelwall & Vaughan, 2004).
 

In Dialog is a command called "rank": "The RANK command provides the ability to perform trend or statistical analysis on an existing search set. The Dialog system extracts terms from the specified field(s) in a set of records and lists them in ranked order, with the most highly posted term appearing first." (Dialog Pocket guide, 2005). (See also Zoom).

 

 

 

Literature:

 

Dialog Pocket guide: http://support.dialog.com/searchaids/dialog/pocketguide/


Meghabghab, G. (2001). Google's web page ranking applied to different topological web graph structures. Journal of the American Society for Information Science and Technology, 52(9), 736-747.
 

Thelwall, M. & Vaughan, L. (2004). New versions of PageRank employing alternative Web document models. ASLIB Proceedings, 56(1), 24-33.
 

 

See also: Major versus minor descriptors; Match

 

 

Birger Hjørland

Last edited: 17-05-2006

Home