|
|
|||||||||||||||||||||||||||||||||
| Definitions | Initials | To take part | Leading Team | Directory |
| A | B | C | D | E | F | G | H | I | J | K | L | M | N | O | P | Q | R | S | T | U | V | W | X | Y | Z |
Definitions
Directory
Dictionary > Definitions > Google > TrustRank - Rank Trust
TrustRank - Rank Trust
Definition
TrustRank is at the same time the name of an algorithm seeking to reduce the spam in the search engines and a registered trademark by Google in Mars 2005.
Description
On March 16, 2005, Google deposited in the United States the TrustRank mark.
Is TrustRank supposed to replace famous PageRank? Difficult to know for the
moment what Google wants to do of TrustRank.
As of on March 11, 2004, 2 researchers (Zoltan Gyongyi and Hector Garcia-Molina)
of the university of Standford (the United States), that one even from where the
cofounders from Google, Larry Page and Sergey Brin come, write with Jan
Pedersen, a researcher of Yahoo! (a competitor of Google) an article entitled
“Combating Web Spam with TrustRank”. This article describes an algorithm which
aims to reduce the spam in the search engines, i.e. to improve the relevance of
the results.
The algorithm of TrustRank proposes a semi-automatic method of classification of
the Web pages (spam or not spam). It is based on a list of pages of reference
guaranteed “clean” after a manual analysis made by human. By propagation and
analyzes bonds between pages, the algorithm determines if the other pages are
also “clean” or if on the contrary they are spam, on the basis of the principle
which the “clean” pages very seldom make bonds towards pages of spam.
The algorithm requires a very good selection of the sites of reference (a simple
list of 200 sites gave good performances according to the article).
To final, it allots a note to each page, its TrustRank (TR). Ranging between 0
(spam) and 1 (page of reference), it indicates the degree of confidence which
one can allot to the page.
TrustRank can be used to only filter the index of an search engine, or in
complement it PageRank to help to classify the results of a research.
Google uses perhaps already this algorithm (or an alternative) since months… or
envisages to do it soon. Difficult to be categorical in this field.
Author
showkat uddin
Author Email
a.zahurul@yahoo.com
References
Definitions to be consulted