Title: The mathematics of Google
Speaker: Professor Michael Brin
Speaker Info: University of Maryland
The order of the documents on the list which is returned by a search engine to the user in response to a query is extremely important. A typical list may contain tens of thousands of web pages, but at best only the first several dozen may be reviewed by the user. Google uses two characteristics of the web page to determine the order of the returned pages - the relevance of the document to the query and the PageRank of the web page. The relevance is based on the relative position, fontification and frequency of the keyword(s) in the document. This factor by itself often does not produce good search results.Date: Wednesday, June 05, 2002
Google uses Markov chains and invariant measures to determine the PageRank of web pages. This, to a large extent, eliminates irrelevant results but also creates problems.
About 2/3 of the talk will be devoted to the general overview of Google and 1/3 to the mathematical methods used in the ordering of the returned list.