Using AI for ranking search results
PageRank is still at the core, but the results have less weight because of the intro-
duction of machine learning technology into ranking, the so-called RankBrain.
According to some sources (see
https://www.bloomberg.com/news/articles/
2015-10-26/google-turning-its-lucrative-web-search-over-to-ai-
machines
), the machine learning algorithm now examines all Google queries and
directly handles 15 percent of the volume of search queries it receives every day,
specializing in
»
Ambiguous and unclear search queries
»
Queries expressed in slang or colloquial terms
»
Queries expressed as though they were occurring in a conversation with the
search engine
Even though RankBrain is still cloaked in secrecy, the algorithm seems to be capable
of guessing, with much higher accuracy than performed by a human being, whether
the contents of a page can appear in search results. It replaces all other ranking fac-
tors in cases that are difficult to judge. This is another example of an additional
algorithm that limits the role played by the original PageRank algorithm.
Do'stlaringiz bilan baham: |