In probability theory and directional statistics, a wrapped asymmetric Laplace distribution is a wrapped probability distribution that results from the "wrapping" of the asymmetric Laplace distribution around the unit circle. For the symmetric case (asymmetry parameter κ = 1), the distribution becomes a wrapped Laplace distribution. The distribution of the ratio of two circular variates (Z) from two different wrapped exponential distributions will have a wrapped asymmetric Laplace distribution. These distributions find application in stochastic modelling of financial data.
The probability density function of the wrapped asymmetric Laplace distribution is:[1]
\begin{align} fWAL(\theta;m,λ,\kappa) &
infty | |
=\sum | |
k=-infty |
fAL(\theta+2\pik,m,λ,\kappa)\\[10pt] &=\dfrac{\kappaλ}{\kappa2+1} \begin{cases} \dfrac{e-(\theta-m)λ\kappa
where
fAL
0\le\theta<2\pi
λ>0
\kappa>0
The cumulative distribution function
FWAL
FWAL(\theta;m,λ,\kappa)=\dfrac{\kappaλ}{\kappa2+1} \begin{cases} \dfrac{emλ\kappa(1-e-\thetaλ\kappa)}{λ\kappa(e2\piλ\kappa-1)}+\dfrac{\kappae-mλ/\kappa(1-e\thetaλ/\kappa)}{λ(e-2\piλ/\kappa-1)}&if\theta\leqm\\ \dfrac{1-e-(\theta-m)λ\kappa
The characteristic function of the wrapped asymmetric Laplace is just the characteristic function of the asymmetric Laplace function evaluated at integer arguments:
\varphi | ||||
|
which yields an alternate expression for the wrapped asymmetric Laplace PDF in terms of the circular variable z=ei(θ-m) valid for all real θ and m:
\begin{align} fWAL(z;m,λ,\kappa) &=
1 | |
2\pi |
infty | |
\sum | |
n=-infty |
-n | |
\varphi | |
n(0,λ,\kappa)z |
\\[10pt] &=
λ | |
\pi(\kappa+1/\kappa) |
\begin{cases} rm{Im}\left(\Phi(z,1,-iλ\kappa)-\Phi\left(z,1,iλ/\kappa\right)\right)-
1 | |
2\pi |
&ifz\ne1 \\[12pt] \coth(\piλ\kappa)+\coth(\piλ/\kappa) &ifz=1 \end{cases} \end{align}
where
\Phi
In terms of the circular variable
z=ei\theta
\langle
n\rangle=\varphi | |
z | |
n(m,λ,\kappa) |
The first moment is then the average value of z, also known as the mean resultant, or mean resultant vector:
\langlez\rangle =
λ2ei | |
\left(1-iλ/\kappa\right)\left(1+iλ\kappa\right) |
The mean angle is
(-\pi\le\langle\theta\rangle\leq\pi)
\langle\theta\rangle=\arg(\langlez\rangle)=\arg(ei)
and the length of the mean resultant is
R=|\langlez\rangle|=
λ2 | ||||
|
The circular variance is then 1 − R
If X is a random variate drawn from an asymmetric Laplace distribution (ALD), then
Z=ei
\theta=\arg(Z)+\pi
0<\theta\leq2\pi
Since the ALD is the distribution of the difference of two variates drawn from the exponential distribution, it follows that if Z1 is drawn from a wrapped exponential distribution with mean m1 and rate λ/κ and Z2 is drawn from a wrapped exponential distribution with mean m2 and rate λκ, then Z1/Z2 will be a circular variate drawn from the wrapped ALD with parameters (m1 - m2, λ, κ) and
\theta=\arg(Z1/Z2)+\pi
-\pi<\theta\leq\pi