Goals of Today's Lecture:
Undergraduate definition of E: integral for absolutely continuous X, sum for discrete. But: rvs which are neither absolutely continuous nor discrete.
General definition of E.
A random variable X is simple if we can write
Def'n: For a simple rv X we define
For positive random variables which are not simple we extend our definition by approximation:
Def'n: If
(almost surely,
)
then
Def'n: We call X integrable if
Facts: E is a linear, monotone, positive operator:
Major technical theorems:
Monotone Convergence: If
a.s. and
(which exists a.s.) then
Dominated Convergence: If
and rv X st
a.s. and
rv Y st
with
then
Fatou's Lemma: If
then
Theorem: With this definition of E if X has density
f(x) (even in
say) and Y=g(X) then
Works even if X has density but Y doesn't.
Def'n:
moment (about origin) of a real
rv X is
(provided it exists).
Generally use
for E(X). The
central moment is
Def'n: For an valued rv X is the vector whose entry is E(Xi)(provided all entries exist).
Def'n: The (
)
variance covariance matrix of X is
Moments and probabilities of rare events are closely connected as will
be seen in a number of important probability theorems. Here is one
version of Markov's inequality (one case is Chebyshev's inequality):
The intuition is that if moments are small then large deviations from
average are unlikely.
Theorem: If
are independent and each Xi is
integrable then
is integrable and
Proof: Usual order: simple Xs first, then positive, then integrable.
Suppose each Xi is simple:
General Xi>0: Xi,n is
Xi rounded down to the nearest multiple of 2-n (to
a maximum of n).
Each Xi,n is simple and
are independent. Thus
The general case uses the fact that
we can write each Xi as the difference of its positive and negative
parts:
Lebesgue Integration
Lebesgue integral defined much the same way as E.
Borel function f simple if
If
almost everywhere and f is Borel define
Call a general f integrable if |f| is integrable and
define for integrable f
Remark: Again you must check that you have not changed the definition of f for either of the previous categories of f.
Facts: is a linear, monotone, positive operator:
Each of these facts is proved first for simple functions then for positive functions then for general integrable functions.
Major technical theorems:
Monotone Convergence: If
almost
everywhere and
(which has to exist almost everywhere) then
Dominated Convergence: If
and there
is a Borel function f such that
for almost
all x and a Borel function g such that
with
then f is integrable and
Fatou's Lemma: If
almost everywhere then
Notice the frequent of almost all or almost everywhere in the hypotheses. In our definition of E wherever we require a property of the function we can require it to hold only for a set of whose complement has probability 0. In this case we say the property holds almost surely. For instance the dominated convergence theorem is usually written:
Dominated Convergence: If
almost surely (
often abbreviated to a.s.) and there
is a random variable X such that
a.s. and
a random variable Y such that
almost surely with
then
The hypothesis of almost sure convergence can be weakened; I hope to discuss this later in the course.
Multiple Integration: Lebesgue integrals over defined using Lebesgue measure on . Iterated integrals wrt Lebesgue measure on give same answer.
Theorem[Tonelli]: If
is
Borel and
almost everywhere
then for almost every
the integral
Theorem[Fubini] If
is
Borel and integrable then for almost every
the
integral