Skip to contents

Runs the variational E-step for each new document against the fitted model's fixed global parameters (topic-word matrix, prior mean and covariance), giving out-of-sample topic proportions (cf. stm::fitNewDocuments). The model's topics are held fixed; only each new document's proportions are estimated.

Usage

fit_new_documents(model, newdata)

Arguments

model

A faSTM fit (non-content; for content models the group-marginal topic-word matrix is used, with a warning).

newdata

A faSTM_corpus, quanteda dfm, or document-term matrix. Terms are aligned to the model's vocabulary; out-of-vocabulary terms are dropped.

Value

A new-documents × K matrix of topic proportions.