Saddlepoint approximation method

Template:Short description The saddlepoint approximation method, initially proposed by Daniels (1954) 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).

Definition

If the moment generating function of a random variable $X=\sum _{i=1}^N{X}_i$ is written as $M(t)=E\left[e^{tX}\right]=E\left[e^{t\sum _{i=1}^N{X}_i}\right]$ and the cumulant generating function as $K(t)=\log (M(t))=\sum _{i=1}^N\log E\left[e^{t{X}_i}\right]$ then the saddlepoint approximation to the PDF of the distribution $X$ is defined as:

${\hat{f}}_X(x)=\frac{1}{\sqrt{2\pi K^{″}(\hat{s})}}\exp (K(\hat{s})-\hat{s}x)(1+ℛ)$

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

${\hat{F}}_X(x)=\{\begin{array}{cc}\Phi (\hat{w})+\varphi (\hat{w})(\frac{1}{\hat{w}}-\frac{1}{\hat{u}})& \text{for }x\ne \mu \\ \frac{1}{2}+\frac{K^{‴}(0)}{6\sqrt{2\pi }{K}^{″}(0{)}^{3/2}}& \text{for }x=\mu \end{array}$

where $\hat{s}$ is the solution to $K^{\prime }(\hat{s})=x$, $\hat{w}=\mathrm{sgn}\hat{s}\sqrt{2(\hat{s}x-K(\hat{s}))}$ ,$\hat{u}=\hat{s}\sqrt{K^{″}(\hat{s})}$, and $\Phi (t)$ and $\varphi (t)$ are the cumulative distribution function and the probability density function of a normal distribution, respectively, and $\mu$ is the mean of the random variable $X$:

$\mu \triangleq E\left[X\right]={\int }_{-\mathrm{\infty }}^{+\mathrm{\infty }}x{f}_X(x)\text{d}x=\sum _{i=1}^{N}E\left[X_i\right]=\sum _{i=1}^N{\int }_{-\mathrm{\infty }}^{+\mathrm{\infty }}{x}_i{f}_{X_i}(x_i)\text{d}{x}_i$.

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.

References

• Butler, Ronald W. (2007), Saddlepoint approximations with applications, Cambridge: Cambridge University Press, ISBN 9780521872508
• Daniels, H. E. (1954), "Saddlepoint Approximations in Statistics", The Annals of Mathematical Statistics, 25 (4): 631–650, doi:10.1214/aoms/1177728652
• Daniels, H. E. (1980), "Exact Saddlepoint Approximations", Biometrika, 67 (1): 59–63, doi:10.1093/biomet/67.1.59, JSTOR 2335316
• Lugannani, R.; Rice, S. (1980), "Saddle Point Approximation for the Distribution of the Sum of Independent Random Variables", Advances in Applied Probability, 12 (2): 475–490, doi:10.2307/1426607, JSTOR 1426607, S2CID 124484743
• Reid, N. (1988), "Saddlepoint Methods and Statistical Inference", Statistical Science, 3 (2): 213–227, doi:10.1214/ss/1177012906
• Routledge, R. D.; Tsao, M. (1997), "On the relationship between two asymptotic expansions for the distribution of sample mean and its applications", Annals of Statistics, 25 (5): 2200–2209, doi:10.1214/aos/1069362394