Invariant estimator

Summary

(Learn how and when to remove this message)

In statistics, the concept of being an invariant estimator is a criterion that can be used to compare the properties of different estimators for the same quantity. It is a way of formalising the idea that an estimator should have certain intuitively appealing qualities. Strictly speaking, "invariant" would mean that the estimates themselves are unchanged when both the measurements and the parameters are transformed in a compatible way, but the meaning has been extended to allow the estimates to change in appropriate ways with such transformations.[1] The term equivariant estimator is used in formal mathematical contexts that include a precise description of the relation of the way the estimator changes in response to changes to the dataset and parameterisation: this corresponds to the use of "equivariance" in more general mathematics.

General setting

edit

Background

edit

In statistical inference, there are several approaches to estimation theory that can be used to decide immediately what estimators should be used according to those approaches. For example, ideas from Bayesian inference would lead directly to Bayesian estimators. Similarly, the theory of classical statistical inference can sometimes lead to strong conclusions about what estimator should be used. However, the usefulness of these theories depends on having a fully prescribed statistical model and may also depend on having a relevant loss function to determine the estimator. Thus a Bayesian analysis might be undertaken, leading to a posterior distribution for relevant parameters, but the use of a specific utility or loss function may be unclear. Ideas of invariance can then be applied to the task of summarising the posterior distribution. In other cases, statistical analyses are undertaken without a fully defined statistical model or the classical theory of statistical inference cannot be readily applied because the family of models being considered are not amenable to such treatment. In addition to these cases where general theory does not prescribe an estimator, the concept of invariance of an estimator can be applied when seeking estimators of alternative forms, either for the sake of simplicity of application of the estimator or so that the estimator is robust.

The concept of invariance is sometimes used on its own as a way of choosing between estimators, but this is not necessarily definitive. For example, a requirement of invariance may be incompatible with the requirement that the estimator be mean-unbiased; on the other hand, the criterion of median-unbiasedness is defined in terms of the estimator's sampling distribution and so is invariant under many transformations.

One use of the concept of invariance is where a class or family of estimators is proposed and a particular formulation must be selected amongst these. One procedure is to impose relevant invariance properties and then to find the formulation within this class that has the best properties, leading to what is called the optimal invariant estimator.

Some classes of invariant estimators

edit

There are several types of transformations that are usefully considered when dealing with invariant estimators. Each gives rise to a class of estimators which are invariant to those particular types of transformation.

  • Shift invariance: Notionally, estimates of a location parameter should be invariant to simple shifts of the data values. If all data values are increased by a given amount, the estimate should change by the same amount. When considering estimation using a weighted average, this invariance requirement immediately implies that the weights should sum to one. While the same result is often derived from a requirement for unbiasedness, the use of "invariance" does not require that a mean value exists and makes no use of any probability distribution at all.
  • Scale invariance: Note that this topic about the invariance of the estimator scale parameter not to be confused with the more general scale invariance about the behavior of systems under aggregate properties (in physics).
  • Parameter-transformation invariance: Here, the transformation applies to the parameters alone. The concept here is that essentially the same inference should be made from data and a model involving a parameter θ as would be made from the same data if the model used a parameter φ, where φ is a one-to-one transformation of θ, φ=h(θ). According to this type of invariance, results from transformation-invariant estimators should also be related by φ=h(θ). Maximum likelihood estimators have this property when the transformation is monotonic. Though the asymptotic properties of the estimator might be invariant, the small sample properties can be different, and a specific distribution needs to be derived.[2]
  • Permutation invariance: Where a set of data values can be represented by a statistical model that they are outcomes from independent and identically distributed random variables, it is reasonable to impose the requirement that any estimator of any property of the common distribution should be permutation-invariant: specifically that the estimator, considered as a function of the set of data-values, should not change if items of data are swapped within the dataset.

The combination of permutation invariance and location invariance for estimating a location parameter from an independent and identically distributed dataset using a weighted average implies that the weights should be identical and sum to one. Of course, estimators other than a weighted average may be preferable.

Optimal invariant estimators

edit

Under this setting, we are given a set of measurements   which contains information about an unknown parameter  . The measurements   are modelled as a vector random variable having a probability density function   which depends on a parameter vector  .

The problem is to estimate   given  . The estimate, denoted by  , is a function of the measurements and belongs to a set  . The quality of the result is defined by a loss function   which determines a risk function  . The sets of possible values of  ,  , and   are denoted by  ,  , and  , respectively.

In classification

edit

In statistical classification, the rule which assigns a class to a new data-item can be considered to be a special type of estimator. A number of invariance-type considerations can be brought to bear in formulating prior knowledge for pattern recognition.

Mathematical setting

edit

Definition

edit

An invariant estimator is an estimator which obeys the following two rules:[citation needed]

  1. Principle of Rational Invariance: The action taken in a decision problem should not depend on transformation on the measurement used
  2. Invariance Principle: If two decision problems have the same formal structure (in terms of  ,  ,   and  ), then the same decision rule should be used in each problem.

To define an invariant or equivariant estimator formally, some definitions related to groups of transformations are needed first. Let   denote the set of possible data-samples. A group of transformations of  , to be denoted by  , is a set of (measurable) 1:1 and onto transformations of   into itself, which satisfies the following conditions:

  1. If   and   then  
  2. If   then  , where   (That is, each transformation has an inverse within the group.)
  3.   (i.e. there is an identity transformation  )

Datasets   and   in   are equivalent if   for some  . All the equivalent points form an equivalence class. Such an equivalence class is called an orbit (in  ). The   orbit,  , is the set  . If   consists of a single orbit then   is said to be transitive.

A family of densities   is said to be invariant under the group   if, for every   and   there exists a unique   such that   has density  .   will be denoted  .

If   is invariant under the group   then the loss function   is said to be invariant under   if for every   and   there exists an   such that   for all  . The transformed value   will be denoted by  .

In the above,   is a group of transformations from   to itself and   is a group of transformations from   to itself.

An estimation problem is invariant(equivariant) under   if there exist three groups   as defined above.

For an estimation problem that is invariant under  , estimator   is an invariant estimator under   if, for all   and  ,

 

Properties

edit
  1. The risk function of an invariant estimator,  , is constant on orbits of  . Equivalently   for all   and  .
  2. The risk function of an invariant estimator with transitive   is constant.

For a given problem, the invariant estimator with the lowest risk is termed the "best invariant estimator". Best invariant estimator cannot always be achieved. A special case for which it can be achieved is the case when   is transitive.

Example: Location parameter

edit

Suppose   is a location parameter if the density of   is of the form  . For   and  , the problem is invariant under  . The invariant estimator in this case must satisfy

 

thus it is of the form   ( ).   is transitive on   so the risk does not vary with  : that is,  . The best invariant estimator is the one that brings the risk   to minimum.

In the case that L is the squared error  

Pitman estimator

edit

The estimation problem is that   has density  , where θ is a parameter to be estimated, and where the loss function is  . This problem is invariant with the following (additive) transformation groups:

 
 
 

The best invariant estimator   is the one that minimizes

 

and this is Pitman's estimator (1939).

For the squared error loss case, the result is

 

If   (i.e. a multivariate normal distribution with independent, unit-variance components) then

 

If   (independent components having a Cauchy distribution with scale parameter σ) then  ,. However the result is

 

with

 

References

edit
  1. ^ see section 5.2.1 in Gourieroux, C. and Monfort, A. (1995). Statistics and econometric models, volume 1. Cambridge University Press.
  2. ^ Gouriéroux and Monfort (1995)
  • Berger, James O. (1985). Statistical decision theory and Bayesian Analysis (2nd ed.). New York: Springer-Verlag. ISBN 0-387-96098-8. MR 0804611.[page needed]
  • Freue, Gabriela V. Cohen (2007). "The Pitman estimator of the Cauchy location parameter". Journal of Statistical Planning and Inference. 137 (6): 1900–1913. doi:10.1016/j.jspi.2006.05.002.
  • Pitman, E.J.G. (1939). "The estimation of the location and scale parameters of a continuous population of any given form". Biometrika. 30 (3/4): 391–421. doi:10.1093/biomet/30.3-4.391. JSTOR 2332656.
  • Pitman, E.J.G. (1939). "Tests of Hypotheses Concerning Location and Scale Parameters". Biometrika. 31 (1/2): 200–215. doi:10.1093/biomet/31.1-2.200. JSTOR 2334983.