0

在 R 实现的 LDA 模型中使用 text2vec 包，但我想知道如何将每个文档分配给主题

BELOW HERE is my code:

library(stringr)
library(rword2vec)
library(wordVectors)
#install.packages("text2vec")
library(text2vec)
library(data.table)
library(magrittr)

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+", " ")
}
movie_review_train = prep_fun(movie_review_train)

tokens = movie_review_train[1:1000] %>% 
  tolower %>% 
  word_tokenizer
it = itoken(tokens, progressbar = FALSE)
v = create_vocabulary(it)
v
vectorizer = vocab_vectorizer(v)
t1 = Sys.time()
dtm_train = create_dtm(it, vectorizer)
print(difftime(Sys.time(), t1, units = 'sec'))
dim(dtm_train)
stop_words = c("i", "me", "my", "myself", "we", "our", "ours", "ourselves")
t1 = Sys.time()
v = create_vocabulary(it, stopwords = stop_words)
print(difftime(Sys.time(), t1, units = 'sec'))
pruned_vocab = prune_vocabulary(v, 
                                term_count_min = 10, 
                                doc_proportion_max = 0.5,
                                doc_proportion_min = 0.001)
vectorizer = vocab_vectorizer(pruned_vocab)
# create dtm_train with new pruned vocabulary vectorizer
t1 = Sys.time()
dtm_train  = create_dtm(it, vectorizer)
print(difftime(Sys.time(), t1, units = 'sec'))
dtm_train_l1_norm = normalize(dtm_train, "l1")
tfidf = TfIdf$new()
# fit model to train data and transform train data with fitted model
dtm_train_tfidf = fit_transform(dtm_train, tfidf)

dtm = transform(dtm_train_tfidf, tfidf)
lda_model <-LDA$new(n_topics = ntopics
                    ,doc_topic_prior = alphaprior
                    ,topic_word_prior = deltaprior
)
lda_model$get_top_words(n = 10, topic_number = c(1:5), lambda = 0.3)

在此之后，我想将每个文档分配给相关主题。我正在获取主题下方的术语列表，但我不知道如何映射。

4

1 回答 1

1

文档主题分布 doc_topic_distr 将每个文档投影到主题空间，可以根据 Dmitriy Selivanov 的文档从下面的代码中计算出来（请参阅http://text2vec.org/topic_modeling.html#example6）。

事实上，主题模型的两个重要输出是主题词矩阵和文档主题矩阵。主题词矩阵或主题词分布显示每个主题中的词权重，而文档主题矩阵或文档主题分布显示主题在每个文档中的贡献。

doc_topic_distr = 
   lda_model$fit_transform(x = dtm, n_iter = 1000, 
                      convergence_tol = 0.001, n_check_convergence = 25, 
                      progressbar = FALSE)
于 2018-11-12T04:06:12.713 回答