spherical_dist: Derivatives of the Log Density of spherical distributions

In dannyjameswilliams/truncsm: Truncated Score Matching

Derivatives of the Log Density of spherical distributions

Description

The first and second derivative of the log density of the von-Mises Fisher distribution and Kent distribution

Usage

psi_vmf(par, x, k = NULL)

psi_kent(par, x, k = NULL, b = NULL)

kent(k = NULL, b = NULL)

vmf(k = NULL)

Arguments

par	parameter vector to be estimated
x	random variable x, taken one row at a time
k	concentration parameter (vmf and kent distribution)
b	ovalness parameter (kent distribution)

Details

psi_vmf corresponds to the von-Mises Fisher distribution, where psi_kent corresponds to the Kent distribution. The parameters to be estimated for the von-Mises Fisher distribution are the mean direction mu and concentration parameter k. Whereas for the Kent distribution, the parameters are mu, the major and minor axes gamma1 and gamma2, the concentration parameter k and ovalness parameter b. The parameters enter psi_kent in a long vector.

Value

list containing two elements

f	the evaluation of the first derivative
grad	the evaluation of the second derivative

dannyjameswilliams/truncsm documentation built on Aug. 22, 2022, 3:44 p.m.