This note is based on Malliavin Calculus and Related Topics by David Nualart and is intended solely for personal academic study and should not be regarded as a substitute for the original text.
1 Wiener Chaos Decomposition
We seek a generalized form of Gaussian process, say changing the set of time indices
to a "more abstract" space
.
Suppose
is a real separable Hilbert space with innrt product
and norm
for
.
Isonormal Gaussian process
Definition 1.1.1. (Gaussian process on
or isonormal Gaussian process)
Stochastic process
defined on a complete probability space
is a Gaussian process on
if
is a centered Gaussian family r.v. and for all
, we have
The mapping
can be shown linear (a.s.). As one can expand and compute
using
. Since the bilinearity of inner product
, we have
, and
, thus we have
This shows that
is an linear isometry from
onto a closed subspace of
, since linear isometry preserves completeness,
space is complete, and thus yields
is closed.
Existance of such probability space. The Kolmogrov extention theorem states that given some probability space
, for all
and finite sequence
, then there exists a probability measure on product
-algebra of
if the measure
on
satisfies
(Permutation Invariant) Given arbitrary permutation
,
and measurable sets
, we have
this means the measure persists when input indices and subscripts are jointly permuted.
2. (Marginal Consistency) Given
and measurable sets
, we have
this tells that the measure is consistent when marginalized out some of its indices.
Then we check whether the isonormal Gaussian process matches the requirements of Kolmogrov extention theorem. Arbitrarily choose
, construct the Gram matrix
whose indices
. The Gram matrix
is symmetric and positive-semidefinite, which corresponds to a Gaussian distribution
, and we use this distribution to denote the distribution of
. We check symmetry conditions first. Given permutation
, there is a corresponding permutation matrix
and the permutated distribution on permutated bases is
. To check the equivalence of two Gaussian measures, it suffices to prove the two covariance matrices is identical under the same set of bases. For permuted distribution we have
, this shows that the measure is invariant under permutation. For marginal consistency, it suffices to prove that given
and
, the top left sub-matrix
of the big covariance matrix
corresponds to the very covariance matrix of
, which is obvious. By Kolmogrov extention theorem, such a measure
DOES exists.
We use Kolmogrov extention theorem to check the well-definiteness of isonormal Gaussian process. A core block of the definition is the covariance structure.
Relation to Ito Brownian Motion and Ito Stochastic Integral. Recall that for one-dimensinoal Brownian motion
, we have
, which is symmetric and semidefinite and corresponds to the Cameron-Martin space
, which is defined as
and the inner product is
. Consider the reproducing kernel
, construct
, whose weak derivative is
. Introduce the Ito stochastic integral operator
, Brownian motion
could be expressed as the stochastic integral of identity function
:
therefore one could rewrite
as
. This setting corresponds with the definition of isonormal Gaussian process:
This reveals that
is the isonormal Gaussian process on
, it also behaves as the isometry, which is the well-known Ito isometry:
We seek to decompose
into orthogonal sum of subspaces, as Taylor expansion allows us to rewrite analytic function into series of functions, we look into Hermite polynomials
for similar building blocks in the sense of Ito integration.