Lecture 16 Flashcards

1
Q

Adjacency matrix, from? to?

A

from top (column indexes) to left (line indexes)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Markov property

A

only the most recent state matters to determine the probability of the next state

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Transition (or Markov) matrices

A

(I,J) entry is a transition probability in [0,1], from state J to state I. Columns add up to one.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Markov A, state x0, how to predict?

A

x1 = A.x0
x2 = A.x1…
Power method

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Steady-state?

A

Long-run equilibrium state reached regardless of the current state, converges to steady-state vector x=Ax s.t x is the eigenvector corresponding to eigenvalue λ=1

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Page rank

A

Probability for a website to be visited by a user randomly clicking on links.
Howto:
Matrix with links from one website to another. Then columns are averaged to get the probability to go from one website to another. If no link in a website, numerical issue, we average to every website.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Page rank issue

A

Markov matrix does not guarantee a unique solution if the graph is not connected.
Perron-Frobenius Theorem: if M markov matrix with all positive entries, then M has a unique steady-state x.

M=0.85A + 0.15/n 1, O(n) (as if 85% chance to clic a link / 15% chance to start over from a random page)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly