Yandex Algorithms: Snezhinsk

The Snezhinsk algorithm became the next version of search algorithms after Arzamas, presented by Yandex in November 2009. The new system of the sites ranking began to take into account almost twice as many factors. A new machine learning technology “MatrixNET” was introduced into the work, which performs the ranking of sites taking into account new requests, with periodic self-learning. What factors does MatrixNET consider? Longer texts supersaturated with keywords became less quoted. The introduction of AGS 17 and AGS 30 filters allowed to remove from the top lines of the search websites created solely for selling links on the Internet without any useful content for users. The rating of websites that steal content from other websites has been lowered (the uniqueness of content consideration). The number of factors of regional dependence of information taken into account when searching has increased. The recognition of abbreviations and transliteration processing as well as indexing of foreign websites has been adjusted. Introduction of the Snezhinsk search algorithm has resulted in reaching the top lines of the search for the most useful content for users and filtering useless, malicious, oversaturated with advertising and spam websites.