One sometimes hears of unscrupulous types who seek to improve
their page-rank Google standings by nefarious means. Yet one must realise that
a pagerank is not something that happens to nice sites that a lot of people might
visit; it is a mathematicl construct based on the number of existing links to a site.
The original algorithm was developed by Larry Page in 1996 - while working at Standford
University and hence the name PageRank - and does not refer to a Miss America pageant
of Web pages.
In mathematics, a Markov Chain is a stochastic process wherein the future state of a system is
a probabilistic result of a current state. The elements of a Markov chain
are thus probability snapshots of a system over time.
A Fourier/Grenoble paper illustrates pagerank with the following example. An arrow from
a site to another is a link, which is assumed to be a positive vote for the reliability of the
site. One can finesse these votes, for example, by dividing outgoinglinks by the number
of links from the site, thus creating a weightd vote. One needs to inject assumptions about black
hole pages with no outgoing links.
.png)
.png)
.png)
.png)
No comments:
Post a Comment