R: bigram

library(dplyr)
library(tidytext)
library(janeaustenr)
library(tidyr)
library(igraph)
library(ggplot2)
library(ggraph)
library(readtext)

text <- readtext("out.txt")
text_bigrams <- text %>%
                unnest_tokens(bigram, text, token = "ngrams", n = 2)
text_bigrams
bigrams_separated <- text_bigrams %>%
separate(bigram, c("word1", "word2"), sep = " ")
#
# stopwords default
bigrams_filtered <- bigrams_separated %>%
       filter(!word1 %in% stop_words$word) %>%
       filter(!word2 %in% stop_words$word)
#
# stopwords Indonesia
bigrams_filtered <- bigrams_separated %>%
       filter(!word1 %in% stopwords::stopwords("id", source = "stopwords-iso")) %>%
       filter(!word2 %in% stopwords::stopwords("id", source = "stopwords-iso"))
bigram_counts <- bigrams_filtered %>%
      count(word1, word2, sort = TRUE)
#
bigram_graph <- bigram_counts %>%
  filter(n > 40) %>%
  graph_from_data_frame()
bigram_graph
set.seed(2017)
ggraph(bigram_graph, layout = "fr") +
  geom_edge_link() +
  geom_node_point() +
  geom_node_text(aes(label = name), vjust = 1, hjust = 1)
#
bigrams_united <- bigrams_filtered %>%
  unite(bigram, word1, word2, sep = " ")
bigrams_united
#
bigram_tf_idf <- bigrams_united %>%
   count(doc_id, bigram) %>%
   bind_tf_idf(doc_id, bigram, n) %>%
   arrange(desc(tf_idf))
bigram_tf_idf

# contoh dari austen book
austen_bigrams <- austen_books() %>%
                  unnest_tokens(bigram, text, token = "ngrams", n = 2)
austen_bigrams

austen_bigrams %>%
      count(bigram, sort = TRUE)

library(tidyr)
bigrams_separated <- austen_bigrams %>%
separate(bigram, c("word1", "word2"), sep = " ")
bigrams_filtered <- bigrams_separated %>%
       filter(!word1 %in% stop_words$word) %>%
       filter(!word2 %in% stop_words$word)

# new bigram counts:
bigram_counts <- bigrams_filtered %>%
      count(word1, word2, sort = TRUE)

bigrams_united <- bigrams_filtered %>%
       unite(bigram, word1, word2, sep = " ")
bigrams_united

austen_books() %>%
   unnest_tokens(trigram, text, token = "ngrams", n = 3) %>%
   separate(trigram, c("word1", "word2", "word3"), sep = " ") %>%
   filter(!word1 %in% stop_words$word,
          !word2 %in% stop_words$word,
          !word3 %in% stop_words$word) %>%
   count(word1, word2, word3, sort = TRUE)

bigram_tf_idf <- bigrams_united %>%
   count(book, bigram) %>%
   bind_tf_idf(bigram, book, n) %>%
   arrange(desc(tf_idf))
bigram_tf_idf

Pranala Menarik

• R