Google RankBrain
Google is constantly trying to update and improve their search engine features for users. Recently, Google has begun using a machine learning AI program called RankBrain along with its current PageRank algorithm to deliver optimized search results. RankBrain seeks to answer queries that even the smartest of Google’s employees cannot. As opposed to PageRank’s methods discussed in class, RankBrain bases its results on accuracy and the relevance of the content it is presented with. Approximately 15% of Google daily queries now use RankBrain, and all are searches Google has not seen before.
Both RankBrain and PageRank are portions of Google’s Hummingbird search algorithm. In class we learned about PageRank and the importance of links pointing to a website. In general, a website will be considered stronger if other sites link to it, especially if those sites are also strong. Links are just one of Google’s many signals that go into its Hummingbird algorithm. Google has realized that the PageRank method is very effective with their search, but there are opportunities for improvement upon this method, which is where other signals are measured, like in RankBrain. We learned in chapter 14 about the calculation of PageRank in terms of scaled random walks, which is a technique that Google has used to branch from to improve their search. Unfortunately, Google does not always publish their changes to the public and therefore their algorithm remains slightly ambiguous.
http://www.business2community.com/seo/google-rankbrain-8-important-facts-know-infographic-01687634#gkKiXKjkmxKYpSHK.97