SMS_HW2
Abstract
比較阿拉丁、灰姑娘、美女與野獸、睡美人四本童話故事之情緒分析1. 載入package
library(dplyr)##
## Attaching package: 'dplyr'## The following objects are masked from 'package:stats':
##
## filter, lag## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, unionlibrary(stringr)
require(tidytext)## Loading required package: tidytextlibrary(wordcloud2)
require(data.table)## Loading required package: data.table##
## Attaching package: 'data.table'## The following objects are masked from 'package:dplyr':
##
## between, first, lastrequire(ggplot2)## Loading required package: ggplot2require(reshape2)## Loading required package: reshape2##
## Attaching package: 'reshape2'## The following objects are masked from 'package:data.table':
##
## dcast, meltrequire(wordcloud)## Loading required package: wordcloud## Warning in library(package, lib.loc = lib.loc, character.only = TRUE,
## logical.return = TRUE, : there is no package called 'wordcloud'require(tidyr)## Loading required package: tidyr##
## Attaching package: 'tidyr'## The following object is masked from 'package:reshape2':
##
## smithsrequire(readr)## Loading required package: readrrequire(scales)## Loading required package: scales##
## Attaching package: 'scales'## The following object is masked from 'package:readr':
##
## col_factor1.1 匯入檔案&資料前處理
Cinderalla<-read_file("Cinderella.txt")
C <- strsplit(Cinderalla, "[.]")[[1]]
C <- data.frame(text = C)
C$text<-C$text %>% as.character()
C <- C %>% mutate(linenumber = row_number())%>%unnest_tokens(word, text)
C$book<-c("Cinderalla")Beauty<-read_file("Beauty and the beast.txt")
B <- strsplit(Beauty, "[.]")[[1]]
B <- data.frame(text = B)
B$text<-B$text %>% as.character()
B <- B %>% mutate(linenumber = row_number())%>%unnest_tokens(word, text)
B$book<-c("Beauty and the Beast")Sleep<-read_file("Sleeping beauty in the wood.txt")
S <- strsplit(Sleep, "[.]")[[1]]
S <- data.frame(text = S)
S$text<-S$text %>% as.character()
S <- S %>% mutate(linenumber = row_number())%>%unnest_tokens(word, text)
S$book<-c("Sleeping Beauty in the Wood")Aladdin<-read_file("ALADDIN AND THE WONDERFUL LAMP.txt")
A <- strsplit(Aladdin, "[.]")[[1]]
A <- data.frame(text = A)
A$text<-A$text %>% as.character()
A <- A %>% mutate(linenumber = row_number())%>%unnest_tokens(word, text)
A$book<-c("Aladdin and the Wonderful Lamp")1.2 合併四本書
Fairy_Tales<-rbind(B,C,S,A)2. NRC 字典過濾出“joy”字詞
nrc_joy <- get_sentiments("nrc") %>%
filter(sentiment == "joy")
Fairy_Tales %>%
filter(book == "Beauty and the Beast") %>%
inner_join(nrc_joy) %>%
count(word, sort = TRUE)## Joining, by = "word"## # A tibble: 77 x 2
## word n
## <chr> <int>
## 1 beauty 80
## 2 found 24
## 3 good 19
## 4 happy 11
## 5 splendid 8
## 6 journey 6
## 7 love 6
## 8 promise 6
## 9 glad 5
## 10 save 5
## # ... with 67 more rows3.1 用bing字典比較四本童話故事的情緒分布
library(tidyr)
bing_Fairy_Tales <- Fairy_Tales %>%
inner_join(get_sentiments("bing")) %>%
count(book, index = linenumber %/% 1, sentiment) %>%
spread(sentiment, n, fill = 0) %>%
mutate(sentiment = positive - negative)%>%
mutate(method = "bing")## Joining, by = "word"bing_Fairy_Tales## # A tibble: 526 x 6
## book index negative positive sentiment method
## <chr> <dbl> <dbl> <dbl> <dbl> <chr>
## 1 Aladdin and the Wonderful Lamp 1 4 1 -3 bing
## 2 Aladdin and the Wonderful Lamp 2 2 0 -2 bing
## 3 Aladdin and the Wonderful Lamp 3 1 0 -1 bing
## 4 Aladdin and the Wonderful Lamp 4 1 0 -1 bing
## 5 Aladdin and the Wonderful Lamp 5 2 1 -1 bing
## 6 Aladdin and the Wonderful Lamp 8 1 0 -1 bing
## 7 Aladdin and the Wonderful Lamp 10 1 0 -1 bing
## 8 Aladdin and the Wonderful Lamp 11 1 0 -1 bing
## 9 Aladdin and the Wonderful Lamp 12 1 0 -1 bing
## 10 Aladdin and the Wonderful Lamp 13 0 3 3 bing
## # ... with 516 more rows- 因為文本字數不多所以index跟原本的行數一樣
3.1.1 bing畫圖
library(ggplot2)
ggplot(bing_Fairy_Tales, aes(index, sentiment, fill = book)) +
geom_col(show.legend = FALSE) +
facet_wrap(~book, ncol = 2, scales = "free_x")
- 童話故事最後的結局幾乎都是正面的，尤其是灰姑娘及美女與野獸。(補充：睡美人最後面幾乎都在講王后的結局，因此負面情緒較高)
- Beauty and the Beast 在大約27行-30之間負面情緒很高，主要是變成野獸的時候十分無助，因為他朋友都以為他已經死了
- Sleeping Beauty在第58行時獲得高分的正面情緒，正是王子親完公主時，對公主表達愛慕之意
3.2 用Afinn字典比較四本童話故事的情緒分數
Afinn_Fairy_Tales <- Fairy_Tales %>%
inner_join(get_sentiments("afinn")) %>%
count(book, index = linenumber %/% 1, score) %>%
mutate(sentiment = score*n)%>%
mutate(method = "Afinn")## Joining, by = "word"Afinn_Fairy_Tales## # A tibble: 913 x 6
## book index score n sentiment method
## <chr> <dbl> <int> <int> <int> <chr>
## 1 Aladdin and the Wonderful Lamp 1 -2 2 -4 Afinn
## 2 Aladdin and the Wonderful Lamp 1 2 1 2 Afinn
## 3 Aladdin and the Wonderful Lamp 2 -3 1 -3 Afinn
## 4 Aladdin and the Wonderful Lamp 2 -2 2 -4 Afinn
## 5 Aladdin and the Wonderful Lamp 4 -3 1 -3 Afinn
## 6 Aladdin and the Wonderful Lamp 8 -3 1 -3 Afinn
## 7 Aladdin and the Wonderful Lamp 9 1 1 1 Afinn
## 8 Aladdin and the Wonderful Lamp 11 -2 1 -2 Afinn
## 9 Aladdin and the Wonderful Lamp 12 -1 1 -1 Afinn
## 10 Aladdin and the Wonderful Lamp 13 2 2 4 Afinn
## # ... with 903 more rows3.2.1 Afinn畫圖
#要把每個index對應的sentiments相加不然圖畫出來數值是錯的
Afinn_Fairy_Tales1<-Afinn_Fairy_Tales %>%
group_by(book,index)%>%
summarise(sentiment=sum(sentiment))%>%
mutate(method = "Afinn")
Afinn_Fairy_Tales1%>%
ggplot(., aes(index, sentiment, fill = book)) +
geom_col(show.legend = FALSE) +
facet_wrap(~book, ncol = 2, scales = "free_x")
3.3 用NRC字典比較四本童話故事的情緒分數
NRC_Fairy_Tales <- Fairy_Tales %>%
inner_join(get_sentiments("nrc")) %>%
filter(sentiment %in% c("positive", "negative"))%>%
count(book, index = linenumber %/% 1, sentiment) %>%
spread(sentiment, n, fill = 0) %>%
mutate(sentiment = positive - negative)%>%
mutate(method = "NRC")## Joining, by = "word"NRC_Fairy_Tales## # A tibble: 566 x 6
## book index negative positive sentiment method
## <chr> <dbl> <dbl> <dbl> <dbl> <chr>
## 1 Aladdin and the Wonderful Lamp 1 1 0 -1 NRC
## 2 Aladdin and the Wonderful Lamp 2 1 0 -1 NRC
## 3 Aladdin and the Wonderful Lamp 3 1 1 0 NRC
## 4 Aladdin and the Wonderful Lamp 4 0 1 1 NRC
## 5 Aladdin and the Wonderful Lamp 5 2 2 0 NRC
## 6 Aladdin and the Wonderful Lamp 6 1 1 0 NRC
## 7 Aladdin and the Wonderful Lamp 7 1 2 1 NRC
## 8 Aladdin and the Wonderful Lamp 8 0 2 2 NRC
## 9 Aladdin and the Wonderful Lamp 9 1 1 0 NRC
## 10 Aladdin and the Wonderful Lamp 10 2 1 -1 NRC
## # ... with 556 more rows3.3.1 NRC畫圖
library(ggplot2)
ggplot(NRC_Fairy_Tales, aes(index, sentiment, fill = book)) +
geom_col(show.legend = FALSE) +
facet_wrap(~book, ncol = 2, scales = "free_x")
3.4 用四本童話故事與三本字典比較的結果
book_all <- bind_rows(bing_Fairy_Tales, Afinn_Fairy_Tales1, NRC_Fairy_Tales)
book_all %>%
ggplot(aes(index, sentiment, fill = method)) +
geom_col(show.legend = FALSE) +
facet_grid(method~book, scales = "free_x") +
labs(x = NULL, y = NULL) 
- 結論：可發現三種字典對於童話故事的正負面字詞比較都差不多，但是Afinn字典所計算出的情緒起伏比較極端，可能是因為Afinn是計算score，bing跟NRC是統計negative與positive次數，所以差距較大
4.最常出現的正面字及負面字
bing_word_counts <- Fairy_Tales %>%
inner_join(get_sentiments("bing")) %>%
count(word, sentiment, sort = TRUE) ## Joining, by = "word"bing_word_counts## # A tibble: 463 x 3
## word sentiment n
## <chr> <chr> <int>
## 1 beauty positive 83
## 2 good positive 39
## 3 great positive 36
## 4 beautiful positive 22
## 5 well positive 20
## 6 fell negative 18
## 7 gold positive 17
## 8 like positive 15
## 9 dead negative 14
## 10 fine positive 14
## # ... with 453 more rows5.1 畫圖
bing_word_counts %>%
group_by(sentiment) %>%
top_n(10) %>%
ungroup() %>%
mutate(word = reorder(word, n)) %>%
ggplot(aes(word, n, fill = sentiment)) +
geom_col(show.legend = FALSE) +
facet_wrap(~sentiment, scales = "free_y") +
labs(y = "Contribution to sentiment",
x = NULL) +
coord_flip()## Selecting by n
- 童話故事書常見的正負面字詞，負面字詞常出現兇猛、貧窮、死亡、失去、恐懼；正面字詞則有漂亮、美麗、愛、滿足等
5.2 統計四本書負向字前十名
#睡美人
Sleep_negative<-Fairy_Tales %>%
filter(book=="Sleeping Beauty in the Wood")%>%
inner_join(get_sentiments("bing") %>% filter(sentiment == "negative"))%>%
count(book,word, sort = TRUE)%>%
head(.,10)## Joining, by = "word"#阿拉丁
Aladdin_negative<-Fairy_Tales %>%
filter(book=="Aladdin and the Wonderful Lamp")%>%
inner_join(get_sentiments("bing") %>% filter(sentiment == "negative"))%>%
count(book,word, sort = TRUE)%>%
head(.,10)## Joining, by = "word"#美女與野獸
Beauty_negative<-Fairy_Tales %>%
filter(book=="Beauty and the Beast")%>%
inner_join(get_sentiments("bing") %>% filter(sentiment == "negative"))%>%
count(book,word, sort = TRUE)%>%
head(.,10)## Joining, by = "word"#Cinderalla
Cinderalla_negative<-Fairy_Tales %>%
filter(book=="Cinderalla")%>%
inner_join(get_sentiments("bing") %>% filter(sentiment == "negative"))%>%
count(book,word, sort = TRUE)%>%
head(.,10)## Joining, by = "word"5.2.1 畫圖
all_negative <- bind_rows(Sleep_negative, Beauty_negative, Aladdin_negative, Cinderalla_negative)
all_negative %>%
ggplot(aes(reorder(word, n), n)) +
geom_col(show.legend = FALSE) +
facet_wrap(~book, scales = "free_y", ncol = 2) +
geom_text(aes(label=n, hjust = -0.5))+
labs(x = "字詞", y = NULL) +
theme(text=element_text(size=12)) +
coord_flip()
- 阿拉丁代表字：slave奴隸；睡美人：fell睡著；美女與野獸：野獸本身有許多負面的形容詞
- 負面的字詞幾乎都有談到貧窮、死亡有關，也是在童話故事中常見的背景描述
5.3 統計四本書正向字前十名
#睡美人
Sleep_positive<-Fairy_Tales %>%
filter(book=="Sleeping Beauty in the Wood")%>%
inner_join(get_sentiments("bing") %>% filter(sentiment == "positive"))%>%
count(book,word, sort = TRUE)%>%
head(.,10)## Joining, by = "word"#阿拉丁
Aladdin_positive<-Fairy_Tales %>%
filter(book=="Aladdin and the Wonderful Lamp")%>%
inner_join(get_sentiments("bing") %>% filter(sentiment == "positive"))%>%
count(book,word, sort = TRUE)%>%
head(.,10)## Joining, by = "word"#美女與野獸
Beauty_positive<-Fairy_Tales %>%
filter(book=="Beauty and the Beast")%>%
inner_join(get_sentiments("bing") %>% filter(sentiment == "positive"))%>%
count(book,word, sort = TRUE)%>%
head(.,10)## Joining, by = "word"#Cinderalla
Cinderalla_positive<-Fairy_Tales %>%
filter(book=="Cinderalla")%>%
inner_join(get_sentiments("bing") %>% filter(sentiment == "positive"))%>%
count(book,word, sort = TRUE)%>%
head(.,10)## Joining, by = "word"5.3.1 畫圖
all_positive <- bind_rows(Sleep_positive, Beauty_positive, Aladdin_positive, Cinderalla_positive)
all_positive %>%
ggplot(aes(reorder(word, n), n)) +
geom_col(show.legend = FALSE) +
facet_wrap(~book, scales = "free_y", ncol = 2) +
geom_text(aes(label=n, hjust = 0))+
labs(x = "字詞", y = NULL) +
theme(text=element_text(size=12)) +
coord_flip()
- 正面字幾乎都有beauty,good,great,gold
- 比較特別的字詞例如阿拉丁有出現magic
6.1 正負面字詞比率
bing <- get_sentiments("bing")
wordcounts <- Fairy_Tales %>%
group_by(book) %>%
summarize(words = n())
plot_Fairt<-Fairy_Tales %>%
inner_join(bing) %>%
group_by(book,sentiment) %>%
summarize(count=n() ) %>%
left_join(wordcounts, by = c("book")) %>%
mutate(ratio= count/words)## Joining, by = "word"plot_Fairt## # A tibble: 8 x 5
## # Groups: book [4]
## book sentiment count words ratio
## <chr> <chr> <int> <int> <dbl>
## 1 Aladdin and the Wonderful Lamp negative 156 5317 0.0293
## 2 Aladdin and the Wonderful Lamp positive 150 5317 0.0282
## 3 Beauty and the Beast negative 241 7196 0.0335
## 4 Beauty and the Beast positive 360 7196 0.0500
## 5 Cinderalla negative 48 2475 0.0194
## 6 Cinderalla positive 96 2475 0.0388
## 7 Sleeping Beauty in the Wood negative 85 3654 0.0233
## 8 Sleeping Beauty in the Wood positive 133 3654 0.03646.2 四本童話故事書正負面比率比較
plot_Fairt %>%
ggplot(aes(fill=sentiment,x=book,y=ratio))+
geom_bar(position="dodge",stat="identity") +
theme(text = element_text(size=10),
axis.text.x = element_text(vjust = 0.5, hjust = 0.5, angle = 15))
- 童話故事比較常出現正面字詞，以四本書來看，只有阿拉丁這本書的出現的負面字詞比率較正面字詞高