GloVe (Global Vectors for Word Representation)

GloVe (Global Vectors for Word Representation)

Glove(input_cooccurences = NULL, output_vectors = NULL,
  vocab_input = NULL, model_output = 0, iter_num = 5,
  learn_rate = 0.1, save_squared_grads_file = NULL,
  alpha_weight = 0.75, cutoff = 100, binary_output = 0,
  vectorSize = 10, threads = 1, trace = FALSE)

Arguments

input_cooccurences	a character string specifying the path to the cooccurences text file
output_vectors	a character string specifying the path to the output-vectors-file(s) (the output depending on the binary_output parameter can be a .bin and/or a .txt file)
vocab_input	a character string specifying the path to the vocabulary text file (the output file of the vocabulary_counts function)
model_output	an integer specifying the model-for-word-vector-output (for text output only). [ 0: output all data, for both word and context word vectors, including bias terms; 1: output word vectors, excluding bias terms; 2: output word vectors + context word vectors, excluding bias terms ]
iter_num	an integer specifying the number of training iterations
learn_rate	a float number specifying the learning rate
save_squared_grads_file	either NULL or a character string specifying the location where the save_squared_grads_file data should be saved (accumulated squared gradients)
alpha_weight	a float number specifying the parameter in exponent of the weighting function
cutoff	a number specifying the cutoff parameter of the weighting function
binary_output	an integer specifying the format output of the saved data (0: text, 1: binary, 2: both)
vectorSize	a number specifying the dimension of word vector representations (excluding the bias term)
threads	an integer specifying the number of threads to run in parallel
trace	either TRUE or FALSE. If TRUE information will be printed out

Value

a character string specifying the location of the saved data and the number of the word vectors

References

https://github.com/stanfordnlp/GloVe

http://nlp.stanford.edu/projects/glove/

http://nlp.stanford.edu/pubs/glove.pdf

Examples

# library(GloveR)

# gl = Glove(input_cooccurences = '/data_GloveR/COOCUR_output.bin', 

#            output_vectors = '/data_GloveR/vectors', 

#            vocab_input = '/data_GloveR/VOCAB.txt', 

#            model_output = 2, iter_num = 5, learn_rate = 0.05, save_squared_grads_file = NULL, 

#            alpha_weight = 0.75, cutoff = 10, binary_output = 0, vectorSize = 50, threads = 6, trace = TRUE)