|
|
| Line 1: |
Line 1: |
| {{Unreferenced|date=October 2009}}
| | Adrianne is what you is going to call me but My family and i don't like when guys and women use my full full name. What [http://Www.dailymail.co.uk/home/search.html?sel=site&searchPhrase=I+genuinely I genuinely] like doing is fish getting and I'll be [https://www.google.com/search?hl=en&gl=us&tbm=nws&q=starting&btnI=lucky starting] something else along with it. Filing is my day job at this time but soon I'll possibly be on my own. Vermont needs always been my tough to adapt place and I are blessed with everything that I need here. Go to the group website to find competeing more: http://circuspartypanama.com<br><br>My page; clash of clans hack android ([http://circuspartypanama.com circuspartypanama.com]) |
| | |
| In [[statistics]], the '''Rao–Blackwell theorem''', sometimes referred to as the '''Rao–Blackwell–Kolmogorov theorem''', is a result which characterizes the transformation of an arbitrarily crude [[estimator]] into an estimator that is optimal by the [[mean squared error|mean-squared-error]] criterion or any of a variety of similar criteria.
| |
| | |
| The Rao–Blackwell theorem states that if ''g''(''X'') is any kind of [[estimator]] of a parameter θ, then the [[conditional expectation]] of ''g''(''X'') given ''T''(''X''), where ''T'' is a [[sufficient statistic]], is typically a better estimator of θ, and is never worse. Sometimes one can very easily construct a very crude estimator ''g''(''X''), and then evaluate that conditional expected value to get an estimator that is in various senses optimal.
| |
| | |
| The theorem is named after [[Calyampudi Radhakrishna Rao]] and [[David Blackwell]]. The process of transforming an estimator using the Rao–Blackwell theorem is sometimes called '''Rao–Blackwellization'''. The transformed [[estimator]] is called the '''Rao–Blackwell estimator'''.
| |
| | |
| ==Definitions==
| |
| *An [[estimator]] δ(''X'') is an ''observable'' random variable (i.e. a [[statistic]]) used for estimating some ''unobservable'' quantity. For example, one may be unable to observe the average height of ''all'' male students at the University of X, but one may observe the heights of a random sample of 40 of them. The average height of those 40—the "sample average"—may be used as an estimator of the unobservable "population average".
| |
| | |
| *A [[sufficiency (statistics)|sufficient statistic]] ''T''(''X'') is a statistic calculated from data ''X'' to estimate some parameter θ for which it is true that no other statistic which can be calculated from data X provides any additional information about θ. It is defined as an ''observable'' [[random variable]] such that the [[conditional probability]] distribution of all observable data ''X'' given ''T''(''X'') does not depend on the ''unobservable'' parameter θ, such as the mean or standard deviation of the whole population from which the data ''X'' was taken. In the most frequently cited examples, the "unobservable" quantities are parameters that parametrize a known family of [[probability distribution]]s according to which the data are distributed.
| |
| ::In other words, a [[sufficiency (statistics)|sufficient statistic]] ''T(X)'' for a parameter θ is a [[statistic]] such that the [[conditional probability|conditional distribution]] of the data ''X'', given ''T''(''X''), does not depend on the parameter θ.
| |
| | |
| *A '''Rao–Blackwell estimator''' δ<sub>1</sub>(''X'') of an unobservable quantity θ is the conditional [[expected value]] E(δ(''X'') | ''T''(''X'')) of some estimator δ(''X'') given a sufficient statistic ''T''(''X''). Call δ(''X'') the '''"original estimator"''' and δ<sub>1</sub>(''X'') the '''"improved estimator"'''. It is important that the improved estimator be ''observable'', i.e. that it not depend on θ. Generally, the conditional expected value of one function of these data given another function of these data ''does'' depend on θ, but the very definition of sufficiency given above entails that this one does not.
| |
| | |
| *The ''[[mean squared error]]'' of an estimator is the expected value of the square of its deviation from the unobservable quantity being estimated.
| |
| | |
| ==The theorem== | |
| | |
| ===Mean-squared-error version===
| |
| One case of Rao–Blackwell theorem states:
| |
| | |
| :The mean squared error of the Rao–Blackwell estimator does not exceed that of the original estimator.
| |
| | |
| In other words
| |
| | |
| :<math>\operatorname{E}((\delta_1(X)-\theta)^2)\leq
| |
| \operatorname{E}((\delta(X)-\theta)^2).\,\!</math>
| |
| | |
| The essential tools of the proof besides the definition above are the [[law of total expectation]] and the fact that for any random variable ''Y'', E(''Y''<sup>2</sup>) cannot be less than [E(''Y'')]<sup>2</sup>. That inequality is a case of [[Jensen's inequality]], although it may also be shown to follow instantly from the frequently mentioned fact that
| |
| | |
| :<math> 0 \leq \operatorname{Var}(Y) = \operatorname{E}((Y-\operatorname{E}(Y))^2) =
| |
| \operatorname{E}(Y^2)-(\operatorname{E}(Y))^2.\,\!</math>
| |
| | |
| ===Convex loss generalization=== | |
| The more general version of the Rao–Blackwell theorem speaks of the "expected loss" or [[risk function]]:
| |
| | |
| :<math>\operatorname{E}(L(\delta_1(X)))\leq \operatorname{E}(L(\delta(X)))\,\!</math>
| |
| | |
| where the "loss function" ''L'' may be any [[convex function]]. For the proof of the more general version, Jensen's inequality cannot be dispensed with.
| |
| | |
| ==Properties==
| |
| The improved estimator is [[bias of an estimator|unbiased]] if and only if the original estimator is unbiased, as may be seen at once by using the [[law of total expectation]]. The theorem holds regardless of whether biased or unbiased estimators are used.
| |
| | |
| The theorem seems very weak: it says only that the Rao–Blackwell estimator is no worse than the original estimator. In practice, however, the improvement is often enormous.
| |
| | |
| ==Example==
| |
| Phone calls arrive at a switchboard according to a [[Poisson process]] at an average rate of λ per minute. This rate is not observable, but the numbers ''X''<sub>1</sub>, ..., ''X''<sub>''n''</sub> of phone calls that arrived during ''n'' successive one-minute periods are observed. It is desired to estimate the probability ''e''<sup>−λ</sup> that the next one-minute period passes with no phone calls.
| |
| | |
| An ''extremely'' crude estimator of the desired probability is
| |
| | |
| :<math>\delta_0=\left\{\begin{matrix}1 & \text{if}\ X_1=0, \\
| |
| 0 & \text{otherwise,}\end{matrix}\right.</math>
| |
| | |
| i.e., it estimates this probability to be 1 if no phone calls arrived in the first minute and zero otherwise. Despite the apparent limitations of this estimator, the result given by its Rao–Blackwellization is a very good estimator.
| |
| | |
| The sum
| |
| | |
| :<math> S_n = \sum_{i=1}^n X_{i} = X_1+\cdots+X_n\,\!</math>
| |
| | |
| can be readily shown to be a sufficient statistic for λ, i.e., the ''conditional'' distribution of the data ''X''<sub>1</sub>, ..., ''X''<sub>''n''</sub>, depends on λ only through this sum. Therefore, we find the Rao–Blackwell estimator
| |
| | |
| :<math>\delta_1=\operatorname{E}(\delta_0\mid S_n=s_n).\,\!</math> | |
| | |
| After doing some algebra we have
| |
| | |
| :<math> | |
| \begin{align}
| |
| \delta_1 & = \operatorname{E}(\operatorname{1}_{\{X_1=0\}} \mid \sum_{i=1}^n X_{i} = s_n)=P(X_{1}=0 \mid \sum_{i=1}^n X_{i} = s_n) \\
| |
| & = \frac{P(X_{1}=0, \sum_{i=2}^n X_{i} = s_n)}{P(\sum_{i=1}^n X_{i} = s_n)} = e^{-\lambda}\frac{\left((n-1)\lambda\right)^{s_n}e^{-(n-1)\lambda}}{s_n!} \Bigg / \frac{(n\lambda)^{s_n}e^{-n\lambda}}{s_n!} = \left(1-{1 \over n}\right)^{s_n}.
| |
| \end{align}
| |
| </math>
| |
| | |
| Since the average number of calls arriving during the first ''n'' minutes is ''n''λ, one might not be surprised if this estimator has a fairly high probability (if ''n'' is big) of being close to
| |
| | |
| :<math>\left(1-{1 \over n}\right)^{n\lambda}\approx e^{-\lambda}.</math>
| |
| | |
| So δ<sub>1</sub> is clearly a very much improved estimator of that last quantity. In fact, since ''S''<sub>''n''</sub> is [[completeness (statistics)|complete]] and δ<sub>0</sub> is unbiased, δ<sub>1</sub> is the unique minimum variance unbiased estimator by the [[Lehmann–Scheffé theorem]].
| |
| | |
| ==Idempotence==
| |
| The Rao–Blackwell process is [[idempotent]]. Using it to improve the already improved estimator does not obtain a further improvement, but merely returns as its output the same improved estimator.
| |
| | |
| ==Completeness and Lehmann–Scheffé minimum variance==
| |
| If the conditioning statistic is both complete and sufficient, and the starting estimator is unbiased, then the Rao–Blackwell estimator is the unique "[[minimum-variance unbiased estimator|best unbiased estimator]]": see [[Lehmann–Scheffé theorem]].
| |
| | |
| ==See also==
| |
| * [[Basu's theorem]] — Another result on complete sufficient and ancillary statistics
| |
| | |
| ==References==
| |
| {{Refbegin}}
| |
| * {{Cite journal
| |
| |last=Blackwell |first=D. |authorlink=David Blackwell
| |
| |title=Conditional expectation and unbiased sequential estimation
| |
| |journal=[[Annals of Mathematical Statistics]]
| |
| |volume=18 |issue=1 |pages=105–110 |year=1947
| |
| |doi=10.1214/aoms/1177730497 |mr=19903 | zbl = 0033.07603
| |
| }}
| |
| *{{Cite journal
| |
| |last=Kolmogorov |first=A. N. |authorlink=Andrey Kolmogorov
| |
| |title=Unbiased estimates
| |
| |journal=Izvestiya Akad. Nauk SSSR. Ser. Mat.
| |
| |year=1950 |volume=14 |issue= |pages=303–326
| |
| |mr=36479
| |
| }}
| |
| {{Refend}}
| |
| | |
| ==External links==
| |
| * {{SpringerEOM |title=Rao–Blackwell–Kolmogorov theorem |id=R/r077550 |first=M.S. |last=Nikulin}}
| |
| | |
| {{DEFAULTSORT:Rao-Blackwell Theorem}}
| |
| [[Category:Statistical theorems]]
| |
| [[Category:Estimation theory]]
| |
Adrianne is what you is going to call me but My family and i don't like when guys and women use my full full name. What I genuinely like doing is fish getting and I'll be starting something else along with it. Filing is my day job at this time but soon I'll possibly be on my own. Vermont needs always been my tough to adapt place and I are blessed with everything that I need here. Go to the group website to find competeing more: http://circuspartypanama.com
My page; clash of clans hack android (circuspartypanama.com)