Uniform integrability

Measure theoretic definition

Textbooks on real analysis and measure theory often use the following definition.

Let ( X , M , μ ) be a positive measure space. A set Φ ⊂ L 1 ( μ ) is called uniformly integrable if to each ϵ > 0 there corresponds a δ > 0 such that

∫ E | f | d μ < ϵ

whenever f ∈ Φ and μ ( E ) < δ .

Probability definition

In the theory of probability, the following definition applies.

A class C of random variables is called uniformly integrable (UI) if given ϵ > 0 , there exists K ∈ [ 0 , ∞ ) such that E ( | X | I | X | ≥ K ) ≤ ϵ    for all X ∈ C , where I | X | ≥ K is the indicator function I | X | ≥ K = { 1 if  | X | ≥ K , 0 if  | X | < K . .

An alternative definition involving two clauses may be presented as follows: A class C of random variables is called uniformly integrable if:

There exists a finite M such that, for every X in C , E ( | X | ) ≤ M and

For every ϵ > 0 there exists δ > 0 such that, for every measurable A such that P ( A ) ≤ δ and every X in C , E ( | X | : A ) ≤ ϵ .

The two probabilistic definitions are equivalent.

Relationship between definitions

The two definitions are closely related. A probability space is a measure space with total measure 1. A random variable is a real-valued measurable function on this space, and the expectation of a random variable is defined as the integral of this function with respect to the probability measure. Specifically,

Let ( Ω , F , P ) be a probability space. Let the random variable X be a real-valued F -measurable function. Then the expectation of X is defined by

provided that the integral exists.

Then the alternative probabilistic definition above can be rewritten in measure theoretic terms as: A set C of real-valued functions is called uniformly integrable if:

There exists a finite M such that, for every X in C , ∫ Ω | X | d P ≤ M .

For every ϵ > 0 there exists δ > 0 such that, for every measurable A such that P ( A ) ≤ δ and for every X in C , ∫ A | X | d P ≤ ϵ .

Comparison of this definition with the measure theoretic definition given above shows that the measure theoretic definition requires only that each function be in L 1 ( μ ) . In other words, ∫ X f d μ is finite for each f , but there is not necessarily an upper bound to the values of these integrals. In contrast, the probabilistic definition requires that the integrals have an upper bound.

One consequence of this is that uniformly integrable random variables (under the probabilistic definition) are tight. That is, for each ϵ > 0 , there exists a > 0 such that

for all X .

In contrast, uniformly integrable functions (under the measure theoretic definition) are not necessarily tight.

In his book, Bass uses the term uniformly absolutely continuous to refer to sets of random variables (or functions) which satisfy the second clause of the alternative definition. However, this definition does not require each of the functions to have a finite integral.

Related corollaries

The following results apply to the probabilistic definition.

Definition 1 could be rewritten by taking the limits as

A non-UI sequence. Let Ω = [ 0 , 1 ] ⊂ R , and define

Clearly X n ∈ L 1 , and indeed E ( | X n | ) = 1   , for all n. However, E ( | X n | , | X n | ≥ K ) = 1    for all  n ≥ K , and comparing with definition 1, it is seen that the sequence is not uniformly integrable.

By using Definition 2 in the above example, it can be seen that the first clause is satisfied as L 1 norm of all X n s are 1 i.e., bounded. But the second clause does not hold as given any δ positive, there is an interval ( 0 , 1 / n ) with measure less than δ and E [ | X m | : ( 0 , 1 / n ) ] = 1 for all m ≥ n .

If X is a UI random variable, by splitting

and bounding each of the two, it can be seen that a uniformly integrable random variable is always bounded in L 1 .

If any sequence of random variables X n is dominated by an integrable, non-negative Y : that is, for all ω and n,

then the class C of random variables { X n } is uniformly integrable.

A class of random variables bounded in L p ( p > 1 ) is uniformly integrable.

Relevant theorems

Dunford–Pettis theorem

A class of random variables X n ⊂ L 1 ( μ ) is uniformly integrable if and only if it is relatively compact for the weak topology σ ( L 1 , L ∞ ) .

de la Vallée-Poussin theorem

The family { X α } α ∈ A ⊂ L 1 ( μ ) is uniformly integrable if and only if there exists a non-negative increasing convex function G ( t ) such that lim t → ∞ G ( t ) t = ∞ and sup α E ( G ( | X α | ) ) < ∞ .

Relation to convergence of random variables

A sequence { X n } converges to X in the L 1 norm if and only if it converges in measure to X and it is uniformly integrable. In probability terms, a sequence of random variables converging in probability also converge in the mean if and only if they are uniformly integrable. This is a generalization of Lebesgue's dominated convergence theorem, see Vitali convergence theorem.