2021/04/24 期中考 N094020026 何明信

#安裝需要的包

packages = c("dplyr", "tidytext", "jiebaR", "gutenbergr", "stringr", "wordcloud2", "ggplot2", "tidyr", "scales",'magrittr ')
existing = as.character(installed.packages()[,1])
for(pkg in packages[!(packages %in% existing)]) install.packages(pkg)

Warning in install.packages :
  package ‘magrittr ’ is not available for this version of R

A version of this package for your version of R might be available elsewhere,
see the ideas at
https://cran.r-project.org/doc/manuals/r-patched/R-admin.html#Installing-packages

#載入

require(dplyr)
require(tidytext)
require(jiebaR)
require(gutenbergr)
require(stringr)
require(wordcloud2)
require(ggplot2)
require(tidyr)
require(scales)
require(magrittr)

#至Gutenberg網站取得編號514，書名Little Women

# 下載 "Anne of Green Gables" 書籍，並且將text欄位為空的行給清除，以及將重複的語句清除
women <- gutenberg_download(514,mirror = "http://mirrors.xmission.com/gutenberg/") %>%
  filter(text!="") %>%
  distinct(gutenberg_id, text) %>%
  mutate(chapter = cumsum(str_detect(text, regex("^CHAPTER.*"))))
head(women,50)

#1-1 >(參考第一章程式碼) 對該書進行斷章與斷詞，算出該書最常用的前十個字，使用 ggplot2 畫出

jieba_tokenizer <- worker()

women_tokenizer <- function(t) {
  lapply(t, function(x) {
    tokens <- segment(x, jieba_tokenizer)
    return(tokens)
  })
}
filter<-c("the","and","you")

tokens_women <- women %>% unnest_tokens(word, text, token=women_tokenizer) #%>%

tokens_women

tokens_count <- tokens_women %>% 
  filter(nchar(.$word)>2) %>%
  group_by(word) %>% 
  summarise(sum = n()) %>% 
  top_n(10, sum) %>%
  arrange(desc(sum)) %>%
  ggplot(aes(x=reorder(word, sum),y=sum)) +
  geom_col() +
  xlab(NULL) +
  coord_flip()

tokens_count

#1-2 >排除停用字(stopwords)後，算出該書最常用的前十個字，使用ggplot2畫出

dir(show_dictpath())

[1] "/Library/Frameworks/R.framework/Versions/4.0/Resources/library/jiebaRD/dict"
[1] "backup.rda"      "hmm_model.zip"   "idf.zip"         "jieba.dict.zip" 
[5] "model.rda"       "README.md"       "stop_words.utf8" "user.dict.utf8" 

jieba_tokenizer <- worker(stop_word = "stop_words.utf8")
women2_tokenizer <- function(t) {
  lapply(t, function(x) {
    tokens <- segment(x, jieba_tokenizer)
    return(tokens)
  })
}
filter<-c("the","and","you")

tokens_women2 <- women %>% unnest_tokens(word, text, token=women2_tokenizer) #%>%
#filter_segment(filter) %>%
#str() 
tokens_women2

tokens_count2 <- tokens_women2 %>% 
  filter(nchar(.$word)>2) %>%
  group_by(word) %>% 
  summarise(sum = n()) %>% 
  top_n(10, sum) %>%
  arrange(desc(sum)) %>%
  ggplot(aes(x=reorder(word, sum),y=sum)) +
  geom_col() +
  xlab(NULL) +
  coord_flip()

tokens_count2

#2 >使用AFINN情緒字典，並以Chapter為單位畫出該書各章節的情緒變化

library(tidytext)

afinn = get_sentiments("afinn")

get_sentiments("afinn")

sentiment_women<-sentiment_count %>%
  group_by(chapter) %>% 
  summarise(sentiment_sum=sum(count))
sentiment_women

sentiment_women_plot<-sentiment_women %>%
  ggplot(aes(x=reorder(chapter, chapter),y=sentiment_sum)) +
  geom_bar(position = 'identity', stat = "identity" , fill="#ffc080")
sentiment_women_plot

#3 >以chapter做為document，計算該書bigram的TF-IDF，去除出現在整本書中只出現過1次的bigram之後，以TF-IDF由大到小列出

tokens_women3<- tokens_women %>%
  filter(nchar(.$word)>1) %>%
  group_by(word,chapter) %>% 
  count()%>%
  bind_tf_idf(word,chapter , n)%>%
  arrange(desc(tf_idf))  
tokens_women3