Notes for A Class of Statistics with Asymptotically Normal Distribution by Wassily Hoeffding.
You can view the lecture notes on U-statistics.
- Basic Concepts and Motivating Examples
- U-statistic
- Asymptotic Normality of U-statistic
- Applications to particular statistics
- References
Basic Concepts and Motivating Examples
Functional, Kernel and Order
Suppose that
Make inference about
Note:
To use all the samples, we can use the U-statistic
where
Here,
Symmetry of Kernel
For convenience, we assume that
But it is not necessary to assume the symmetry of the kernel. Because we can always convert a non-symmetric kernel to a symmetric one.
If
where the sum is over all permutations of
Motivating Examples
Example 1: Let
Example 2: Let
Example 3: Let
Asymptotic Normality of U-statistic: A Simple Version
Define
where
U-statistic
Notations
: independent random vectors with the same d.f. . : -dimensional random vector. : sample of -dimensional vectors. : a symmetric function of vector arguments. : functional of .
Definition: U-statistic
Consider the function of the sample,
where the kernel
Eq (4.4) can be driven from
where
Asymptotic Normality of U-statistic
The Variance of a U-statistic
The Unbiased Estimator and Its Variance
If
Let
where
and
Define
We have
Suppose that the variance of
We have
Stationary Order of a Functional
If, for some parent distribution
If
The Variance of a U-statistic: i.i.d. Case
If
If the variance of
where
and exactly
are satisfied. By (5.12), each term in
and hence, since
The Variance of a U-statistic: General Case
When the distributions of
where the sum is extended over all subscripts
Then the variance of
Properties of the Moments and the Variance
Returning to the case of identically distributed
Theorem 5.1 The quantities
Theorem 5.2 The variance
which takes on its upper bound
If
where
A Necessary and Sufficient Condition for the Existence of the Variance
(5.13) and (5.19) imply that a necessary and sufficient condition for the existence of
or that of
If
If
Lemma 5.1
For proving Theorem 5.1 we shall require the following:
Lemma 5.1. If
we have
and
Proof. (5.29) follows from (5.27) by induction.
For proving (5.28) let
Then, by (5.10),
and on substituting this in (5.27) we have
From (5.9) it is seen that (5.28) is true for
Let
For an arbitrary fixed
Then, by induction hypothesis,
for any fixed
Now,
and hence
The proof of Lemma 5.1 is complete.
Proof of Theorems 5.1
By (5.29) we have for
From (5.28), and since
Proof of Theorem 5.2.
From (5.19) we have
Applying these inequalities to each term in (5.13) and using the identity
we obtain (5.20).
(5.22) and (5.23) follow immediately from (5.13).
For (5.21) we may write
where
Let
Then we have from (5.13)
or
Putting
where
Hence, by (5.19),
and
By (5.33) and (5.31), the latter sum vanishes. This proves (5.32).
For the stationary case
Properties of the Covariance
Let’s talk about the covariance of two U-statistics. Consider a set of
each
Let
If, in particular,
Let
be the covariance of
In a similar way as for the variance, we find, if
The right hand side is easily seen to be symmetric in
For
We have from (5.23) and (6.5)
Hence, if
The Limit Theorems: i.i.d. Case
In this section the vectors
Notes:
Converge of the Distribution Function:
A sequence of d.f.’s
Singularity of the Distribution:
A
LEMMA 7.1. Let
Then the d.f. of
The Limit Theorem 7.1 and 7.2
Theorem 7.1. Let
Let
be
where the summation is over all subscripts such that
and
exist, the joint d.f. of
tends, as
According to Theorem 5.2,
Theorem 7.2. Under the conditions of Theorem 7.1, and if
the joint d.f. of
tends, as
Proof of Theorem 7.1. The existence of (7.4) entails that of
which, by (5.19), (5.20) and (6.6), is sufficient for the existence of
Now, consider the
where
By the Central Limit Theorem for vectors (cf. Cramer [1, p. 112]), the joint d.f. of
Theorem 7.1 will be proved by showing that the
have the same joint limiting distribution as
According to Lemma 7.1 it is sufficient to show that
For proving (7.7), write
By (5.13) we have
and from (7.5),
By (7.2) and (6.1) we may write for (7.6)
and hence
The term
is
and 0 otherwise. For a fixed
On inserting (7.9), (7.10), and (7.12) in (7.8), we see that (7.7) is true.
The proof of Theorem 7.1 is complete.
The Limit Theorem 7.3: Extension Theorem 7.1 to a Larger Class of Statistics
The application of Lemma 7.1 leads immediately to the following extension of Theorem 7.1 to a larger class of statistics.
Theorem 7.3. Let
where
tends to the normal distribution with zero means and covariance matrix
The Limit Theorem 7.4: Application to Sample Functionals
The theorem 7.3 applies, in particular, to the regular functionals
in the case that the variance of
where the sum
where the expected value
Theorem 7.4. Let
be
exists, the joint d.f. of
tends to the g-variate normal d.f. with zero means and covariance matrix
The Limit Theorem 7.5: Application to Functions of Statistics
The following theorem is concerned with the asymptotic distribution of a function of statistics of the form
Theorem 7.5. Let
Applications to particular statistics
- Moments and functions of moments
- Mean different and coefficient of concentration
- Functions of ranks and of the signs of variate differences
- Difference sign correlation
- Rank correlation and grade correlation
- Non-parametric tests of independence
- Mann’s test against trend
- The coefficient of partial difference sign correlation
References
[1] W. Hoeffding, ‘A Class of Statistics with Asymptotically Normal Distribution’, Ann. Math. Statist., vol. 19, no. 3, pp. 293–325, Sep. 1948, doi: 10.1214/aoms/1177730196.
[2] Shao J. Mathematical statistics[M]. Springer Science & Business Media, 2003.
v1.5.2