Google BrainRank, A Human Solution to A Human Problem
https://www.techgenyz.com/2018/02/09/google-rankbrain-seo-need-know/
RankBrain is a web search algorithm first launched by Google in 2015 and now gradually spreads its application in almost every query entry into Google. It particularly targets words that have never been queried before, by assigning a mathematical vector to each keyword and mapping out the relationship between the unknown search words and related words in mathematical space, in which way RankBrain effectively guesses the meaning the unknown words and returns the best-match results.
Unlike linked-based page ranking, which prioritises webpages most endorsed by authorities who ensemble better lists via summing up incoming links based on network structure, Google RankBrain resorts to a “human solution to solve a human problem”. It evaluates how accurately the delivered search results resemble what browsers really want depending on what search links browsers click on. The most frequently visited will be ranked up when someone else searches a similar term, while ranking of unvisited connections will be lowered. The visits also help this artificial intelligence to examine how relative the determined meaning of keywords is and adjust its assigned mathematical vector. RankBrain demonstrates great potential to improve accuracy of search results in general, as Google encounters massive amounts of unknown queries thanks to versatility of human language. However, because RankBrain is not a natural language processor—–in other words, it cannot interpret the meaning of keywords on its own but to find a closest relative of existing words, RankBrain requires a sufficiently large and diverse database still based on the standard content or link based web search.