Column-wise MLE of the angular Gaussian and the von Mises Fisher distributions {Directional}

Column-wise MLE of the angular Gaussian and the von Mises Fisher distributions

Description

Column-wise MLE of the angular Gaussian and the von Mises Fisher distributions.

Usage

colspml.mle(x ,tol = 1e-07, maxiters = 100, parallel = FALSE)
colvm.mle(x, tol = 1e-07)

Arguments

x

A numerical matrix with data. Each column refers to a different vector of observations of the same distribution. The values of for Lognormal must be greater than zero, for the logitnormal they must by percentages, exluding 0 and 1, whereas for the Borel distribution the x must contain integer values greater than 1.

tol

The tolerance value to terminate the Newton-Raphson algorithm.

maxiters

The maximum number of iterations that can take place in each regression.

parallel

Do you want this to be executed in parallel or not. The parallel takes place in C++, and the number of threads is defined by each system's availiable cores.

Details

For each column, spml.mle function is applied that fits the angular Gaussian distribution estimates its parameters and computes the maximum log-likelihood.

Value

A matrix with four columns. The first two are the mean vector, then the \gamma parameter, and the fourth column contains maximum log-likelihood.

Author(s)

Michail Tsagris.

R implementation and documentation: Michail Tsagris [e-mail].

References

Presnell Brett, Morrison Scott P. and Littell Ramon C. (1998). Projected multivariate linear models for directional data. Journal of the American Statistical Association, 93(443): 1068–1077.

See Also

spml.mle, spml.reg, vmf.mle

Examples

x <- matrix( runif(100 * 10), ncol = 10)
a <- colspml.mle(x)
b <- colvm.mle(x)
x <- NULL