This action might not be possible to undo. Are you sure you want to continue?

BooksAudiobooksComicsSheet Music### Categories

### Categories

### Categories

### Publishers

Scribd Selects Books

Hand-picked favorites from

our editors

our editors

Scribd Selects Audiobooks

Hand-picked favorites from

our editors

our editors

Scribd Selects Comics

Hand-picked favorites from

our editors

our editors

Scribd Selects Sheet Music

Hand-picked favorites from

our editors

our editors

Top Books

What's trending, bestsellers,

award-winners & more

award-winners & more

Top Audiobooks

What's trending, bestsellers,

award-winners & more

award-winners & more

Top Comics

What's trending, bestsellers,

award-winners & more

award-winners & more

Top Sheet Music

What's trending, bestsellers,

award-winners & more

award-winners & more

P. 1

Markov Chains Norris|Views: 613|Likes: 10

Published by Genie Kcl

See more

See less

https://www.scribd.com/doc/82262886/Markov-Chains-Norris

08/19/2013

text

original

A set I is countable if there is a bijection f : {I, ... ,n} ~ I for some n E N,-

or a bijection f : N ~ I. In either case we can enumerate all the elements

of I

218

6. Appendix: probability and measure

where in one case the sequence terminates and in the other it does not.

There would have been no loss in generality had we insisted that all our

Markov chains had state-space N or {I, ... ,n} for some n E N: this just

corresponds to a particular choice of the bijection f.

Any subset of a countable set is countable. Any finite cartesian product

of countable sets is countable, for example tln

for any n. Any countable

union of countable sets is countable. The set of all subsets of N is uncount-

able and so is the set of real numbers JR.

We need the following basic fact.

** Lemma 6.1.1.** Let I be a countably infinite set and let Ai ~ 0 for all i E I.

Then, for any two enumerations ofI

~ 1 , ~ 2 , ~ 3 , ••• ,

we have

00

00

LAin = LAin'

n=l n=l

Proof. Given any N E N we can find M ~ Nand N' ~ M such that

Then

N

M

N'

""" A· < """ A· < """ A·

L.-J 'tn - L.-J In - L.-J 'tn

n=l n=l n=l

and the result follows on letting N ~ 00. D

Since the value of the sum does not depend on the enumeration we are

justified in using a notation which does not specify an enumeration and

write simply

More generally, if we allow Ai to take negative values, then we can set

where

6.1 Countable sets and countable sums

219

allowing that the sum over I is undefined when the sums over I+ and I-

are both infinite. There is no difficulty in showing for Ai, jji 2 0 that

I)Ai + Pi) = LAi + LPi.

iEI

iEI iEI

By induction, for any finite set** J** and for Aij 2 0, we have

L (LAij)= L (LAij).

iEI jEJ

jEJ iEI

The following two results on sums are simple versions of fundamental

results for integrals. We take the opportunity to prove these simple versions

in order to convey some intuition relevant to the general case.

Lemma 6.1.2 (Fubini's theorem - discrete case). Let I and** J** be

countable sets and let Aij 2 0 for all i E I and j E

L (LAij)=L (LAij).

iEI jEJ

jEJ iEI

Proof. Let jl,j2,j3, ... be an enumeration of** J.** Then

as n` ~` 00. Hence

and the result follows by symmetry. D

Lemma 6.1.3 (Monotone convergence - discrete case). Suppose for

each i E I we are given an increasing sequence` ( A i ( n ) ) n ~ O` with limit Ai, .

and that Ai(n) 2 0 for all i and n. Then

LAi(n) i LAi asn-t 00.

iEI

iEI

220

6. Appendix: probability and measure

Proof. Set 8i (1) = Ai(l) and for n` ~` 2 set

Then 8i (n)` ~` 0 for all i and n, so as n` ~` 00, by Fubini's theorem

`~ A i ( n )` =` ~` (t,8i(k))

= t.` ( ~ 8 i ( k ) )` i t.` ( ~ 8 i ( k ) )
`

= L(f8i(k)) = LAio

iEI k=l

iEI

- Read and print without ads
- Download to keep your version
- Edit, email or read offline

Lee, Nelder & Pawitan 2006 - Generalized Linear Models With Random Effects

Markov Chain Random Walk

2008 - Claeskens, Hjort.pdf

Actuarial Mathematics

0521867061 - Andrew Gelman - Data Analysis Using Regression and Multilevel~Hierarchical Models [2006]

Generalized Linear Models 2nd Ed

Stochastic Processes - Ross

MLC

Bayesian Data Analysis

Probabilistic Graphical Models Principles and Techniques

123738050 Applied Linear Statistical Models 5th Ed M Kutner Et Al McGraw Hill Irwin 2005 WW

Actuarial Exams Manuals

time-series-introduction.pdf

Introduction Graph Theory

Applied Linear Statistical Models -Neter Et Al (McGraw Hill Fifth Edition 2005)

Probability Theory and Stochastic Processes with Applications

Cohen Y. - Statistics and Data with R - 2008.pdf

Adventures in Stochastic Processes (Sidney Resnick)(2)

David R. Brillinger Time Series Data Analysis and Theory 2001

Regression Analysis of Count Data

Research Design and Statistical Analysis

CT8-PC-09

Nonparametric Statistical Methods Using R

temporal data mining

Linear Regression

Linear Algebra and Matrices

gubner

Clustering for Data Mining

Understanding Advanced Statistical Methods(PDF)

Finite Markov chains and algorithmic applications

Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

CANCEL

OK

You've been reading!

NO, THANKS

OK

scribd

/*********** DO NOT ALTER ANYTHING BELOW THIS LINE ! ************/ var s_code=s.t();if(s_code)document.write(s_code)//-->