estim: Estimate Poisson parameter matrix with reduced rank.

In chroetz/poisrrr: Poisson Reduced-Rank Models

Description

Estimate Poisson parameter matrix with reduced rank.

Usage

estim(Y, K, verbose = TRUE, tol = 1e-05, max_iter = 100, return_list = FALSE)

Arguments

Y	A n x d matrix containing token counts for n different tokens and d different documents.
K	A positive integer. The rank of the normalized parameter matrix.
verbose	A logical value. Should progress messages be printed on the console?
tol	A positive double. Stop iterative optimization of log-likelihood (ll) if improvement is less than tol*ll.
max_iter	A positive integer. Stop iterative optimization of log-likelihood after max_iter iterations.
return_list	A logical value or a numeric vector. Should a list of parameter matrices for each iteration (or all iterations with number in return_list) be returned or just the last?

Value

A n x d matrix or a list of such matrices. A matrix contains the logs of the Poisson parameters.

chroetz/poisrrr documentation built on Dec. 19, 2021, 4:05 p.m.