pairwise_sample.R

tar %>% 
  mutate(speech_id = 1:nrow(.)) %>% 
  unnest_tokens(senten, content, 
                token="sentences") %>% 
  select(speech_id, senten) %>% 
  mutate(sentence_id = 1:nrow(.)) %>% 
  unnest_tokens(morph, senten,
                token = pos) %>% 
  filter(str_detect(morph, "/n|/v(v|a)")) %>% 
  mutate(
    morph = if_else(
      str_detect(morph, "/n"),
      str_replace_all(morph, "/.*$", ""),
      str_replace_all(morph, "/.*$", "다")
    )
  ) %>% 
  filter(str_length(morph) > 1) %>% 
  pairwise_count(morph, sentence_id, sort=T, upper=F) %>% 
  filter(item1 == "사랑")