Created
November 12, 2017 20:42
-
-
Save mwufi/79b04582629508fd826688ec4822daf7 to your computer and use it in GitHub Desktop.
Create vocabulary and document-term matrix using text2vec
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
get_dtm <- function(movie_review){ | |
setDT(movie_review) | |
setkey(movie_review, id) | |
# vectorization | |
prep_fun = function(x) { | |
x %>% | |
# make text lower case | |
str_to_lower %>% | |
# remove non-alphanumeric symbols | |
str_replace_all("[^[:alpha:]]", " ") %>% | |
# collapse multiple spaces | |
str_replace_all("\\s+", " ") | |
} | |
tok_fun = word_tokenizer | |
it_train = itoken(movie_review$body, | |
preprocessor = prep_fun, | |
tokenizer = tok_fun, | |
ids=movie_review$id, | |
progressbar = FALSE) | |
# Start the clock! | |
ptm <- proc.time() | |
vocab = create_vocabulary(it_train) | |
vocab = vocab %>% prune_vocabulary(term_count_min = 10, | |
doc_proportion_max = 0.5) | |
# Stop the clock | |
time <- proc.time() - ptm | |
print(time) | |
print('Time to create vocabulary') | |
# iterator for the vocabulary | |
vectorizer = vocab_vectorizer(vocab) | |
# Start the clock! | |
ptm <- proc.time() | |
dtm_train = create_dtm(it_train, vectorizer) | |
# Stop the clock | |
time <- proc.time() - ptm | |
print(time) | |
print('Time to create document-term matrix') | |
return(list(dtm=dtm_train, vocab=vocab, iterator=it_train)) | |
} | |
# user system elapsed | |
# 50.480 0.248 50.745 | |
# [1] "Time to create vocabulary" | |
# user system elapsed | |
# 53.244 0.340 53.589 | |
# [1] "Time to create document-term matrix" | |
stuff <- get_dtm(t) | |
dtm <- stuff$dtm | |
vocab <- stuff$vocab | |
it <- stuff$iterator | |
dim(dtm) | |
# [1] 92335 21361 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment