Scribd
Upload
154 views2 pages

Markov Kernel Wikipedia

Markov kernels generalize the concept of a transition matrix to allow for Markov processes with infinite state spaces. A Markov kernel is a map that associates a probability measure to each point in a source space, describing the transition probabilities to a target space. Examples include kernels for simple random walks on integers and Galton-Watson branching processes. Kernels can be estimated using kernel density estimation and characterize the conditional distribution of a random variable given a sub-sigma-algebra.

Uploaded by

FRANCESCO222
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as TXT, PDF, TXT or read online on Scribd
154 views2 pages

Markov Kernel Wikipedia

Markov kernels generalize the concept of a transition matrix to allow for Markov processes with infinite state spaces. A Markov kernel is a map that associates a probability measure to each point in a source space, describing the transition probabilities to a target space. Examples include kernels for simple random walks on integers and Galton-Watson branching processes. Kernels can be estimated using kernel density estimation and characterize the conditional distribution of a random variable given a sub-sigma-algebra.

Uploaded by

FRANCESCO222
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as TXT, PDF, TXT or read online on Scribd
You are on page 1/ 2

Markov kernel

From Wikipedia, the free encyclopedia


(Redirected from Stochastic kernel estimation)
In probability theory, a Markov kernel (or stochastic kernel) is a map that play
s the role, in the general theory of Markov processes, that the transition matri
x does in the theory of Markov processes with a finite state space.[1][2]
Contents
1 Formal definition
2 Examples
3 Properties
3.1 Semidirect product
3.2 Regular conditional distribution
4 Estimation
5 References
Formal definition
Let (X,\mathcal A), (Y,\mathcal B) be measurable spaces. A Markov kernel with so
urce (X,\mathcal A) and target (Y,\mathcal B) is a map \kappa \colon X \times \m
athcal B \to [0,1] with the following properties:
The map x \mapsto \kappa(x,B) is \mathcal A - measureable for every B \in \m
athcal B .
The map B \mapsto \kappa(x,B) is a probability measure on (Y, \mathcal B) fo
r every x \in X.
(i.e. It associates to each point x \in X a probability measure \kappa(x,.) on (
Y,\mathcal B) such that, for every measurable set B\in\mathcal B, the map x\maps
to \kappa(x,B) is measurable with respect to the \sigma-algebra \mathcal A.)
Examples
Simple random walk: Take X=Y=\Z and \mathcal A = \mathcal B = \mathcal P(\Z)
, then the Markov kernel \kappa with
\kappa(x,B)=\frac{1}{2}\delta_{x-1}(B)+\frac{1}{2}\delta_{x+1}(B), \quad \fo
rall x \in \Z \forall B \in \mathcal P(\Z),
describes the transition rule for the random walk on \Z.
Galton-Watson process: Take S=Y=\N, \mathcal A = \mathcal B = \mathcal P(\N)
, then
\kappa(x,B)=\begin{cases} \delta_0(B) & \quad x=0,\\ P[\xi_1 + \dots + \xi_x
\in B] & \quad \text{else,}\\ \end{cases}
with i.i.d. random variables \xi_i.
General Markov processes with finite state space: Take X=Y, \mathcal A = \ma
thcal B = \mathcal P(X) = \mathcal P(Y) and |X|=|Y|=n, then the transition rule
can be represented as a stochastic matrix (K_{ij})_{1 \leq i,j \leq n} with \Sig
ma_{j \in Y}K_{ij}=1 for every i \in X. In the convention of Markov kernels we w
rite
\kappa(i,B)=\Sigma_{j \in B}K_{ij} \quad \forall i \in X \forall B \in \math
cal B.
Properties

Semidirect product
Let (X, \mathcal A, P) be a probability space and \kappa a Markov kernel from (X
, \mathcal A) to some (Y, \mathcal B). Then there exists a unique measure Q on (
X \times Y, \mathcal A \otimes \mathcal B), s.t.
Q(A \times B) = \int_A \kappa(x,B)dP(x), \quad \forall A \in \mathcal A \for
all B \in \mathcal B.
Regular conditional distribution
Let (S,Y) be a Borel space, X a (S,Y) - valued random variable on the measure sp
ace (\Omega, \mathcal F,P) and \mathcal G \subseteq \mathcal F a sub-\sigma-alge
bra. Then there exists a Markov kernel \kappa from (\Omega, \mathcal G) to (S,Y)
, s.t. \kappa(.,B) is a version of the conditional expectation E[\mathbf 1_{\{X
\in B\}}| \mathcal G] for every B \in Y, i.e.
P[X \in B|\mathcal G]=E[\mathbf 1_{\{X \in B\}}|\mathcal G]=\kappa(\omega,B)
, \quad P-a.s. \forall B \in \mathcal G.
It is called regular conditional distribution of X given \mathcal G and is not u
niquely defined.
Estimation
The kernel can be estimated using kernel density estimation.[3]
References
Epstein, P.; Howlett, P.; Schulze, M. S. (2003). "Distribution dynamics: Str
atification, polarization, and convergence among OECD economies, 1870 1992". Explo
rations in Economic History 40: 78. doi:10.1016/S0014-4983(02)00023-2. edit
Reiss, R. D. (1993). "A Course on Point Processes". Springer Series in Stati
stics. doi:10.1007/978-1-4613-9308-5. ISBN 978-1-4613-9310-8. edit
Poletti Laurini, M. R.; Valls Pereira, P. L. (2009). "Conditional stochastic
kernel estimation by nonparametric methods". Economics Letters 105 (3): 234. do
i:10.1016/j.econlet.2009.08.012. edit
Bauer, Heinz (1996), Probability Theory, de Gruyter, ISBN 3-11-013935-9
36. Kernels and semigroups of kernels

You might also like