Please check the latest news (change log) and keep this package updated.
wordvec
as the primary class of word vectors
data: Now the data classes contain wordvec
,
data.table
, and data.frame
, which actually
perform as a data.table
.train_wordvec()
function: Train word vectors using
the Word2Vec, GloVe, or FastText algorithm
with multi-threading.tokenize()
function: Tokenize raw texts for
training word vectors.data_wordvec_reshape()
function: Reshape word
vectors data from dense (a data.table
of new classs
wordvec
with two variables word
and
vec
) to plain (a matrix
of word vectors) or
vice versa.test_RND()
function, and tab_WEAT()
is
renamed to test_WEAT()
: These two functions serve as
convenient tools of word semantic similarity analysis and conceptual
association test.plot_wordvec_tSNE()
function: Visualize 2-D or 3-D
word vectors with dimensionality reduced using the t-Distributed
Stochastic Neighbor Embedding (t-SNE) method.data_wordvec_subset()
function.unique
argument for
tab_similarity()
.test_WEAT()
.