# Edgeworth series

Edgeworth series

The Gram-Charlier A series and the Edgeworth series, named in honor of Francis Ysidro Edgeworth, are series that approximate a probability distribution in terms of its cumulants. The series are the same; but, the arrangement of terms (and thus the accuracy of truncating the series) differ.

Gram-Charlier A series

The key idea of these expansions is to write the characteristic function of the distribution whose probability density function is "F" to be approximated in terms of the characteristic function of a distribution with known and suitable properties, and to recover "F" through the inverse Fourier transform.

Let "f" be the characteristic function of the distribution whose density function is "F", and &kappa;"r" its cumulants. We expand in terms of a known distribution with probability density function $Psi$, characteristic function $psi$, and standardized cumulants &gamma;"r". The density $Psi$ is generally chosen to be that of the normal distribution, but other choices are possible as well. By the definition of the cumulants, we have the following formal identity:

:$f\left(t\right)=expleft \left[sum_\left\{r=1\right\}^infty\left(kappa_r-gamma_r\right)frac\left\{\left(it\right)^r\right\}\left\{r!\right\} ight\right] psi\left(t\right),.$

By the properties of the Fourier transform, ("it")"r"&psi;("t") is the Fourier transform of (−1)"r" "D""r" $Psi$("x"), where "D" is the differential operator with respect to "x". Thus, we find for "F" the formal expansion

:$F\left(x\right) = expleft \left[sum_\left\{r=1\right\}^infty\left(kappa_r - gamma_r\right)frac\left\{\left(-D\right)^r\right\}\left\{r!\right\} ight\right] Psi\left(x\right),.$

If $Psi$ is chosen as the normal density with mean and variance as given by "F", that is, mean &mu; = &kappa;1 and variance &sigma;2 = &kappa;2, then the expansion becomes

:$F\left(x\right) = expleft \left[sum_\left\{r=3\right\}^inftykappa_rfrac\left\{\left(-D\right)^r\right\}\left\{r!\right\} ight\right] frac\left\{1\right\}\left\{sqrt\left\{2pi\right\}sigma\right\}expleft \left[-frac\left\{\left(x-mu\right)^2\right\}\left\{2sigma^2\right\} ight\right] ,.$

By expanding the exponential and collecting terms according to the order of the derivatives, we arrive at the Gram-Charlier A series. If we include only the first two correction terms to the normal distribution, we obtain

:$F\left(x\right) = frac\left\{1\right\}\left\{sqrt\left\{2pi\right\}sigma\right\}expleft \left[-frac\left\{\left(x-mu\right)^2\right\}\left\{2sigma^2\right\} ight\right] left \left[1+frac\left\{kappa_3\right\}\left\{3!sigma^3\right\}H_3left\left(frac\left\{x-mu\right\}\left\{sigma\right\} ight\right)+frac\left\{kappa_4\right\}\left\{4!sigma^4\right\}H_4left\left(frac\left\{x-mu\right\}\left\{sigma\right\} ight\right) ight\right] ,,$

with "H"3("x") = "x"3 − 3"x" and "H"4("x") = "x"4 − 6"x"2 + 3 (these are Hermite polynomials).

Note that this expression is not guaranteed to be positive, and is therefore not a valid probability distribution. The Gram-Charlier A series diverges in many cases of interest&mdash;it converges only if "F"("x") falls off faster than exp(−"x"2/4) at infinity (Cramér 1957). When it does not converge, the series is also not a true asymptotic expansion, because it is not possible to estimate the error of the expansion. For this reason, the Edgeworth series (see next section) is generally preferred over the Gram-Charlier A series.

Edgeworth series

Edgeworth developed a similar expansion as an improvement to the central limit theorem. The advantage of the Edgeworth series is that the error is controlled, so that it is a true asymptotic expansion.

Let "X""i" be a sequence of independent identically distributed random variables, and "Y""n" the standardized sum

:$Y_n = frac\left\{sum_\left\{i=1\right\}^n\left(X_i-mbox\left\{E\right\} \left[X_i\right] \right)\right\}\left\{sqrt\left\{sum_\left\{i=1\right\}^nmbox\left\{var\right\} \left[X_i\right] .$

Further, let "F""n" be the probability density function of the variables "Y""n". By the central limit theorem,

:$lim_\left\{n ightarrowinfty\right\} F_n\left(x\right) = frac\left\{1\right\}\left\{sqrt\left\{2piexp\left(- frac\left\{1\right\}\left\{2\right\}x^2\right)$

for every "x", as long as the means and variances are finite and the sum of variances diverges to infinity. (Generally, the conclusion of the central limit theorem is about the limit of cumulative distribution functions, not of probability density functions, and therefore applies to discrete distributions as well. But discrete distributions are not contemplated in the present context).

Now assume that the random variables "X""i" have mean &mu;, variance &sigma;2, and higher cumulants &kappa;"r"=&sigma;"r"&lambda;"r". If we expand in terms of the unit normal distribution, that is, if we set

:$Psi\left(x\right)=frac\left\{1\right\}\left\{sqrt\left\{2piexp\left(- frac\left\{1\right\}\left\{2\right\}x^2\right)$

then the cumulant differences in the formal expression of the characteristic function "f""n"(t) of "F""n" are

:$kappa_1-gamma_1 = 0,,$

:$kappa_2-gamma_2 = 0,,$

:$kappa_r-gamma_r = frac\left\{lambda_r\right\}\left\{n^\left\{r/2-1; qquad rgeq 3,.$

The Edgeworth series is developed similarly to the Gram-Charlier A series, only that now terms are collected according to powers of "n". Thus, we have

:$f_n\left(t\right)=left \left[1+sum_\left\{j=1\right\}^infty frac\left\{P_j\left(it\right)\right\}\left\{n^\left\{j/2 ight\right] exp\left(-t^2/2\right),,$

where "P""j"("x") is a polynomial of degree 3"j". Again, after inverse Fourier transform, the density function "F""n" follows as

:$F_n\left(x\right) = Psi\left(x\right) + sum_\left\{j=1\right\}^infty frac\left\{P_j\left(-D\right)\right\}\left\{n^\left\{j/2 Psi\left(x\right),.$

The first three terms of the expansion are (Cramér 1957)

:$F_n\left(x\right)approx Psi\left(x\right) - frac\left\{lambda_3 Psi^\left\{\left(3\right)\right\}\left(x\right)\right\}\left\{6sqrt\left\{n +frac\left\{1\right\}\left\{n\right\}left \left[frac\left\{lambda_4 Psi^\left\{\left(4\right)\right\}\left(x\right)\right\}\left\{24\right\}+frac\left\{lambda_3^2 Psi^\left\{\left(6\right)\right\}\left(x\right)\right\}\left\{72\right\} ight\right] +O\left(1/n^\left\{3/2\right\}\right),.$

Here, $Psi$("j")("x") is the "j"th derivative of $Psi$("x") with respect to "x". Blinnikov and Moessner (1998) have given a simple algorithm to calculate higher-order terms of the expansion.

* H. Cramér (1957). "Mathematical Methods of Statistics". Princeton University Press, Princeton.
* D. L. Wallace (1958). "Asymptotic approximations to distributions". "Ann. Math. Stat." 29:635-654.
* P. McCullagh (1987). "Tensor Methods in Statistics". Chapman and Hall, London.
* D. R. Cox and O. E. Barndorff-Nielsen (1989). "Asymptotic Techniques for Use in Statistics". Chapman and Hall, London.
* P. Hall (1992). "The Bootstrap and Edgeworth Expansion". Springer, New York.
* S. Blinnikov and R. Moessner (1998). "Expansions for nearly Gaussian distributions". "Astron. Astrophys. Suppl. Ser." 130:193-205.
* J. E. Kolassa (2006). "Series Approximation Methods in Statistics, 3rd Edition". (Lecture Notes in Statistics #88). Springer, New York.

Wikimedia Foundation. 2010.

### Look at other dictionaries:

• Edgeworth — may refer to:;People * Francis Ysidro Edgeworth (1845 ndash;1926), Irish statistician and economist * Henry Essex Edgeworth (1745 ndash;1807), Catholic priest and confessor of Louis XVI * Maria Edgeworth (1767 ndash;1849), novelist * Richard… …   Wikipedia

• EDGEWORTH, MARIA —    novelist, born at Blackbourton, Berks; from her fifteenth year her home was in Ireland; she declined the suit of a Swedish count, and remained till the close of her life unmarried; amongst the best known of her works are Moral Tales, Tales… …   The Nuttall Encyclopaedia

• Edgeworth, Maria — (1767 1849)    Novelist, only child of Richard E., of Edgeworthstown, Co. Longford, was b. near Reading. Her f., who was himself a writer on education and mechanics, bestowed much attention on her education. She showed early promise of… …   Short biographical dictionary of English literature

• Francis Ysidro Edgeworth — (8 February 1845 ndash; 13 February 1926) made significant contributions to the methods of statistics during the 1880s. From 1891 onward he was the editor of a leading academic journal in economics and his own writings in economics were… …   Wikipedia

• List of characters in the Ace Attorney series — Contents 1 Recurring characters 1.1 Phoenix Wright 1.2 Miles Edgeworth …   Wikipedia

• Ace Attorney Investigations: Miles Edgeworth — Developer(s) Capcom Publisher(s) Capcom, THQ (Australia) Director(s) …   Wikipedia

• Maria Edgeworth — Maria Edgeworth, ca. 1841 Born 1 January 1768(1768 01 01) Black Bourton, Oxfordshire, England Died …   Wikipedia

• Discography of the Ace Attorney series — The Ace Attorney series of video games has had many soundtracks. Contents 1 Gyakuten Saiban Yomigaeru Gyakuten Original Soundtrack 2 Gyakuten Saiban + Gyakuten Saiban 2 Original Soundtrack 3 Gyakuten Saiba …   Wikipedia

• Belinda (Edgeworth novel) — Infobox Book | name = Belinda title orig = translator = image caption = author = Maria Edgeworth illustrator = cover artist = country = United Kingdom language = English series = genre = Novel publisher = Joseph Johnson pub date = 1801 english… …   Wikipedia

• Marvel vs. Capcom (series) — This article is about the series. For the game, see Marvel vs. Capcom: Clash of Super Heroes. Marvel vs. Capcom (マーヴルVSカプコン, Māburu tai Kapukon?) is a series of fighting games created by Capcom in which characters created by Marvel Comics and… …   Wikipedia