No replies to this topic

[DOTOANNANG]

The calculation of PageRanks from the Google matrix relies on several deep theorems of mathematics:

1. Since $G$ is a positive stochastic matrix (that is, every entry is positive and each row sums to $1$), an important theorem of linear algebra, proved by Oskar Perron in $\it{1907}$, guarantees that there is a unique positive probability vector $\bf{v}$ that satisfies ${\bf{v}}= {\bf{v}}G$.
2. The theory of Markov chains guarantees to produce PageRank scores with approximately $m$ digits of accuracy, the number of iterations should be around $m/\lg\left ( 1/d \right )$.

Edited by DOTOANNANG, 24-10-2023 - 15:15.