Chapter 5 Perron Frobenius Theorem Flashcards
What is the adjacency matrix
0, 1 matrix
What is the H matrix
Obtained by normalising the rows of the adjacency matrix and so is non negative with sums of 1 on rows
State the perron frobenius theorem
A is n x n with all entries non negative then:
- A had real eigenvlaue p, p>=|lamda| for every eigenvalue of A
- There is eigenvector v with non negative entries and row eigenvector W ^transpose corresponding the the eigenvalue p
- If A had positive entries P is simple. furthermore p>|lamda| for all eigenvalues lamda not qual to p of A
Why do we have the perron frobenius theorem
For really big matrices we cannot necessarily find the jordan blocks or eigenvalues to tell me about the limit
What does a simple eigenvalue mean
It is not repeated
If A is positive then how is p simple?
As eigenvalue of A and vectors v and w have positive entries
What is the name given to eigenvalue p in this theorem
perron eigenvalues
What is the name given to the nonnegative eigenvector v
Perron eigenvector
What is a key thing to remember when finding the left and right perron eigenvectors
They are positive typically so we should choose their values accordingly. Can be non negative if A is also non negative.
What is a stochastic matrix
An n x n matrix is stochastic if all entries are non negative and each row sum is 1
How does the perron frobenius theorem relate to stochastic matrix
Then 1 is the perron eigenvalue of S and e = (1,1,1….1) is the corresponding perron eigenvector and we have |lamda|<=1 for all other eigenvalues
What does column stochastic mean
Matrix A is column stochastic if all entries are non negative and each column sum of A is 1. 1 is a left perron eigenvector of a column stochastic A
What is doubly stochastic matrix
Non negative matrix that is both stochastic and column stochastic
What does perron frobenius tell us about a stochastic matrix with all positive entries
Perron eigenvalue is 1 and all other eigenvalues are less than 1. It has right eigenvector e= (1,1,1,…,1)
When does the limit exist for a stochastic matrix
Only when it is positive as then only 1 eigenvalue is equal to 1