Mathematics at Google Brief history • Started in 1996 as the research project ‘Backrub’ by the then PhD student Larry Page • Sergey Brin joined in • Became incorporated in 1998 • Began selling advertisement in 2000 • Became public in 2004 – Initial raise $2,718,281,828 (e * 10 ^ 9) – Implied value of 23 billion The 25,000,000,000 vector • Represent the web as a huge directed graph • Constantly traversed by a “random web surfer” • Forward and backlinks • The importance of each page is based on importance of the pages linking to it • Importance of a page • Transition matrix • Stohastic matrix – All entries are non negative – The sum of all entries in a column is 1 (unless the node corresponding to that column has no links) Dangling nodes • Nodes with no forward links Dangling nodes (continued) • If A is the matrix whose entries are all zero except for the columns corresponding to dangling nodes, in which each entry is 1/n, then S = H + A. Final Matrix • Google Matrix • α – damping factor, usually 0.85 Other examples • Google maps – graph of all cities. - Selected route is the shortest path between the nodes - YouTube rankings - Ad serving - Selection algorithm - Various areas of research Forecasting • Simulation of the ad server • Sampling of data • Ad selection based on satisfaction index of the ad Machine Learning • Marriage between statistics and computer science • Adaptive models for predicting and classifying data regression analysis, neural networks, support vector machines Math in management Chief Economist of Google is Hal Varian - professor in University of Berkeley specializing in microeconomics and information technology - Employee ratings - Resource allocations