Saddlepoint approximation method explained

The saddlepoint approximation method, initially proposed by Daniels (1954)[1] is a specific example of the mathematical saddlepoint technique applied to statistics, in particular to the distribution of the sum of

N

independent random variables. It provides a highly accurate approximation formula for any PDF or probability mass function of a distribution, based on the moment generating function. There is also a formula for the CDF of the distribution, proposed by Lugannani and Rice (1980).[2]

Definition

If the moment generating function of a random variable

X=

N
\sum
i=1

Xi

is written as

M(t)=E\left[etX\right]=

N
t\sumXi
i=1
E\left[e

\right]

and the cumulant generating function as

K(t)=log(M(t))=

N
\sum
i=1

log

tXi
E\left[e

\right]

then the saddlepoint approximation to the PDF of the distribution

X

is defined as:

\hat{f}X(x)=

1
\sqrt{2\piK''(\hat{s

)}}\exp(K(\hat{s})-\hat{s}x)\left(1+l{R}\right)

where

l{R}

contains higher order terms to refine the approximation and the saddlepoint approximation to the CDF is defined as:

\hat{F}X(x)=\begin{cases}\Phi(\hat{w})+\phi(\hat{w})\left(

1
\hat{w
} - \frac\right) & \text x \neq \mu \\ \frac + \frac & \text x = \mu \end where

\hat{s}

is the solution to

K'(\hat{s})=x

,

\hat{w}=sgn{\hat{s}}\sqrt{2(\hat{s}x-K(\hat{s}))}

,

\hat{u}=\hat{s}\sqrt{K''(\hat{s})}

, and

\Phi(t)

and

\phi(t)

are the cumulative distribution function and the probability density function of a normal distribution, respectively, and

\mu

is the mean of the random variable

X

:

\mu\triangleqE\left[X\right]=

+infty
\int
-infty

xfX(x)dx=

N
\sum
i=1

E\left[Xi\right]=

N
\sum
i=1
+infty
\int
-infty

xi

f
Xi

(xi)dxi

.

When the distribution is that of a sample mean, Lugannani and Rice's saddlepoint expansion for the cumulative distribution function

F(x)

may be differentiated to obtain Daniels' saddlepoint expansion for the probability density function

f(x)

(Routledge and Tsao, 1997). This result establishes the derivative of a truncated Lugannani and Rice series as an alternative asymptotic approximation for the density function

f(x)

. Unlike the original saddlepoint approximation for

f(x)

, this alternative approximation in general does not need to be renormalized.

Notes and References

  1. Daniels . H. E. . December 1954. Saddlepoint Approximations in Statistics . The Annals of Mathematical Statistics . en . 25 . 4 . 631–650 . 10.1214/aoms/1177728652 . 0003-4851.
  2. Lugannani . Robert . Rice . Stephen . June 1980 . Saddle point approximation for the distribution of the sum of independent random variables . Advances in Applied Probability . en . 12 . 2 . 475–490 . 10.2307/1426607 . 0001-8678. subscription .