20181013 Verbatim Practice -TM 사용

20181013 Verbatim Practice - TM

if("rJava" %in% installed.packages("rJava") == FALSE)install.packages("rJava")
library(rJava)
if("memoise" %in% installed.packages("memoise") == FALSE)install.packages("memoise")
library(memoise)
if("stringr" %in% installed.packages("stringr") == FALSE)install.packages("stringr")
library(stringr)
if("KoNLP" %in% installed.packages("KoNLP") == FALSE)install.packages("KoNLP")
library(KoNLP)
if("tm" %in% installed.packages("tm") == FALSE)install.packages("tm")
library(tm)
if("wordcloud" %in% installed.packages("wordcloud") == FALSE)install.packages("wordcloud")
library(wordcloud)
if("RColorBrewer" %in% installed.packages("RColorBrewer") == FALSE)install.packages("RColorBrewer")
library(RColorBrewer)
if("dplyr" %in% installed.packages("dplyr") == FALSE)install.packages("dplyr")
library("dplyr")

install.packages("readxl")
library("readxl")

setwd('C:\\Users\\Administrator\\rlang_weekend2\\Data_R_181013')
tgt1  <- read_excel("Verbatim.xlsx", 1)
# 만약 sheet가 여러개이면 숫자만, 2, 3, 4로 변경가능
head(tgt1)

# useSejongDic()
# KoNLP::buildDictionary(
#   ext_dic = c('sejong','woorimalsam')
# )
# 
# verbatim <- str_replace_all(verbatim,"\\W"," ")
# head(verbatim)
# 
# nouns <- sapply(
#                 verbatim,
#                 extractNoun,
#                 USE.NAMES = F
# )
# 
# nouns <- extractNoun(verbatim)
# wordcount <- table(unlist(nouns))
# df_word <- as.data.frame(wordcount,stringAsFactor = F)
# df_word <- dplyr::rename(df_word,
#                          word = Var1,
#                          freq = Freq)
# # head(df_word)
# df_word <- filter(df_word, nchar(word)>=2)
# 
# top20 <- df_word %>% 
#   arranage(desc(freq)) %>% 
#   head(20)

tgt2 <- VCorpus(VectorSource(tgt1))
tgt2 <- tm_map(tgt2, tm::stripWhitespace) # 공백처리
tgt2 <- tm_map(tgt2, tolower) # 알파벳이면 소문자
tgt2 <- tm_map(tgt2, removePunctuation) # 마침표, 공백, 세미콜론, 콜론제거

gsb <- c(stopwords(
  ('english')
))
tgt2 <- tm_map(tgt2, removeWords, gsb)
tgt2 <- tm_map(tgt2, PlainTextDocument)
tgt3 <- TermDocumentMatrix(tgt2)
findFreqTerms(tgt3, 5)
findAssocs(tgt3, "BMW", 0.5)
tgt4 <- as.matrix(tgt3)
head(tgt4)
tgt4 <- sort(rowSums(tgt4),decreasing = T)
tgt4

# KoNLP 사용해서 gsb function 사요

pal <- brewer.pal(8,"Dark2")
set.seed(1234) 
wordcloud(
  names(tgt4),
  freq = tgt4,
  scale = c(2.5,0.1), # 단어크기 0.1 ~ 2.5
  rot.per =0.25, # 회전비율
  min.freq = 2, # 최저 빈도수 2회이상
  random.order = F, # 고빈도 단어 중앙배치
  random.color = T,
  colors = pal
)