von Mises-Fisher mixture model reduction - Greedy method

When given parameters of the von Mises-Fisher mixture model, this function aims at mixture model reduction using a greedy method.

movMF_reduce_greedy(means, concentrations, weights, target.num = 2)

Arguments

means

a \((K \times p)\) matrix of means of the von Mises-Fisher components.

concentrations

a \(K\) vector of concentrations of the von Mises-Fisher components.

weights

a \(K\) vector of weights of the von Mises-Fisher components.

target.num

a desired number of components after reduction. Default is 2.

Value

a named list of the reduced mixture model containing

means

a \((\code{target.num} \times p)\) matrix of means of the von Mises-Fisher components.

concentrations

a \(\code{target.num}\) vector of concentrations of the von Mises-Fisher components.

weights

a \(\code{target.num}\) vector of weights of the von Mises-Fisher components.