as.textmodel_doc2vec: Create a doc2vec model

In wordvector: Word and Document Vector Models

Create a doc2vec model

Description

Create a doc2vec model as weighted word vectors.

Usage

as.textmodel_doc2vec(x, model, compound = TRUE, group_data = FALSE, ...)

Arguments

x	a quanteda::tokens or quanteda::dfm object.
model	a textmodel_wordvector object.
compound	if TRUE, compound multi-word expressions in x based on model internally. Only applies when x is a quanteda::tokens object.
group_data	if TRUE, apply dfm_group(x) before creating document vectors.
...	additional arguments passed to the underlying function.

Details

For Japanese or Chinese texts, model$concatenator must be empty (""). The value is inherited from the tokens object on which the model was trained. It triggers tokenization of words in model and compounding of characters in x before creating document vectors.

Value

Returns a textmodel_doc2vec object with the following elements:

values	a list of matrices for word and document vectors.
dim	the size of the document vectors.
concatenator	the concatenator in x.
docvars	document variables copied from x.
call	the command used to execute the function.
version	the version of the wordvector package.

wordvector documentation built on Oct. 8, 2026, 9:11 a.m.