You are on page 1of 19

4 How google works

4.1 Cluster architecture


The slides were adapted from Jaremy Bradley; used with permission.

33
34
35
36
37
38
39
4.2 The pagerank algorithm

40
41
42
43
44
45
46
47
48
49
50
Exercise

Graphs G1 and G2 represent a set of connected web pages. Taking c = 0.8 write down for each
graph

• the transition matrix P

• the personalisation vector p~

• the modified transition matrix A

• two iterations ~x(k+1) = xk A starting with x0 = (0, 1, 0) for G1 and x0 = (1, 0, 0, 0) for G2 .

• two iterations of the PageRank algorithm for comparison

51

You might also like