Thenmfbin R package provides a simple Non-Negative Matrix Factorization (NMF) implementation tailored for binary data matrices. It offers a choice of initialization methods, loss functions and updating algorithms.

NMF is typically used for reducing high-dimensional matrices into lower (k-) rank ones wherek is chosen by the user. Given a non-negative matrixX of size$m\times n$, NMF looks for two non-negative matricesW ($m\times k$) andH ($k\times n$), such that:

$$X\approx W\times H$$

In topic modelling, ifX is a word-document matrix thenW can be interpreted as the word-topic matrix andH as the topic-document matrix.

Unlike most other NMF packages,nmfbin is focused on binary (Boolean) data, while keeping the number of dependencies to a minimum. For more information see thewebsite.

# install.packages("remotes")remotes::install_github("michalovadek/nmfbin")

# loadlibrary(nmfbin)# Create a binary matrix for demonstrationX<-matrix(sample(c(0,1),100, replace=TRUE), ncol=10)# Perform Logistic NMFresults<-nmfbin(X, k=3, optimizer="mur", init="nndsvd", max_iter=1000)

@Manual{,  title = {nmfbin: Non-Negative Matrix Factorization for Binary Data},  author = {Michal Ovadek},  year = {2023},  note = {R package version 0.2.1},  url = {https://michalovadek.github.io/nmfbin/},}