train_embeddings: Trains embeddings from your corpus using methods described...

In adrianapaza/easynlp:

Description

Trains embeddings from your corpus using methods described here: http://nlp.stanford.edu/pubs/glove.pdf. Using the text2vec package; see that package for more info.

Usage

train_embeddings(vocab, it_all, vocab_vectorizer, window = 10,
  dimensions = 100, max_iters = 50, max_cooccur = 50)

Arguments

it_all	The tokens from Create_Vocab_Document_Term_Matrix.
vocab_vectorizer	The vocabulary vectorizer from Create_Vocab_Document_Term_Matrix
window	The window size for word co-occurences
dimensions	The number of dimensions returned for word embeddings. Defaults to 100
max_iters	The maximum number of iterations for training the embeddings. Defaults to 50
max_cooccur	The maximum number of times a word-word co-occurence may be used in weighting the model. Defaults to 50. Value should be proportional to amount of data.
input	A dataframe of a text corpus

Value

Returns a dataframe of word embeddings

Examples

train_embeddings(Myvocab, itokens, vocab_vectorizer, window=10, dimensions=100, max_iters=50, max_cooccur=50)

adrianapaza/easynlp documentation built on May 9, 2019, 7:31 p.m.