nmfem (version 1.0.5)

nmfem_mult_modelselection: Model selection in NMF-EM algorithm for mixture of multinomials

Description

The function proceed to a model selection with NMF-EM algorithm on mixture of multinomials dataset. The function use the automated dimension jump approach implemented in the capushe package.

Usage

nmfem_mult_modelselection(X, maxK = 30, save = FALSE, path = ".")

Arguments

X

numeric matrix containing multinomials observations of dimension N (number of observation) x M (number of variables).

maxK

integer. Maximum number of clusters to be tested. By default, function tests from 2 to 30 clusters.

save

logical. Whether the result of each parameter couple (H,K) tested got to be saved.

path

path to save the results if save = TRUE. By default, it is the working directory. Three directories are created to save the results. Directory "Initializations" contains the initialization of the algorithm for each value of K. Matrices are saved in directory "Matrices", and plots in directory "Results".

Examples

# NOT RUN {
# Example on the complete data - needs around an hour to run
# }
# NOT RUN {
nmfem_mult_modelselection(travelers[ ,-1])
# }