R语言文本分析《三国演义》

数据读取与包的加载

由于最近对文本分析比较感兴趣，所以分析三国演义的文本。

getwd()
setwd("E:\\三国")
library(jiebaR)
library(ggplot2)
library(jpeg)
library(reshape2)
library(wordcloud)

读取前十行

> sanguo <- readLines("E:/三国/三国演义白话文版.txt")#逐行读取数据
> sanguo[1:10]#展现数据前十行[1] "三国演义明?罗贯中"                                                                     [2] "致读者"                                                                                [3] "　“大江东去，浪淘尽，千古风流人物……”北宋大文学家苏东坡的一首《念奴娇.赤壁》"           [4] "　　，仅用百字，就生动地使三国英雄的形象跃然纸上，再现了三国时火烧赤壁的悲壮惨烈、波澜"[5] "　　壮阔的战争场面。"                                                                  [6] "　　《三国演义》是中国历史上继《水浒传》之后的又一部伟大的现实主义巨著，是中国古典"    [7] "　　文学宝库中的又一灿烂的瑰宝，波澜壮阔，气象万千。《三国演义》全名《三国志通俗演义"  [8] "　　》，取材于三国时近百年的历史事实，经作者进行了文学创作，终成一部浩瀚的鸿篇巨帙，"  [9] "　　流传至今，脍炙人口。东汉末年是诸侯割据、天下大乱的年代，英雄造时势，时势出英雄，"  
[10] "　　政"     

分词处理

> dictpath <- "三国停词.txt"
> stoppath <- "stopwords.dat" #设置停用词
> cutter <- worker(user="E:/三国/三国停词.txt", bylines = TRUE, stop_word="E:/三国/stopwords.dat")#进行分词
res <- cutter[sanguo]#分词 
> head(res)#展示前六行
[[1]]
[1] "三国演义" "明"       "罗贯中"  [[2]]
[1] "读者"[[3]][1] "　"       "大江东去" "浪淘尽"   "千古"     "风流人物" "北宋"    [7] "文学家"   "苏东坡"   "一首"     "念奴娇"   "赤壁"    [[4]][1] "　"       "　"       "仅用"     "百字"     "生动"     "地使"    [7] "三国"     "英雄"     "形象"     "跃然纸上" "再现"     "三国"    
[13] "时"       "火烧"     "赤壁"     "悲壮"     "惨烈"     "波澜"    [[5]]
[1] "　"       "　"       "壮阔"     "战争场面"[[6]][1] "　"       "　"       "三国演义" "中国"     "历史"     "上继"    [7] "水浒传"   "一部"     "现实主义" "巨著"     "中国"     "古典"    

词频统计

> text <- unlist(res)#设置数据类型
> freq <- data.frame(table(text))#设置数据框
> freq <- freq[nchar(as.character(freq$text))>=2,]#提取字符串大于等于2的词
> freq <- freq[order(-freq$Freq),]#对词频进行排序
> str(freq)#浏览freq对象的存储类型和结构
>  head(freq)#读取前六行数据text Freq
1176  曹操 2430
6792  刘备 2266
6130  孔明 1794
3936  关公  862
7023  吕布  692
14075 张飞  602

描绘词频图

roles <- c("曹操|孟德|阿瞒","刘备|玄德|刘玄德","孔明|诸葛亮|卧龙|诸葛孔明","关公|关羽|云长|关云长|长生","吕布|奉先|飞将军","翼德|张飞") #输入人物称谓
role_name = c("曹操","刘备","孔明","关公", "吕布","张飞")#输入人物称谓
role_paras = sapply(roles, grepl, text) #判断
colnames(role_paras) = role_name #修改列名
role_counts = data.frame(role = factor(colnames(role_paras), levels = c("曹操","刘备","孔明","关公", "吕布","张飞")), count = colSums(role_paras)) #统计library(ggplot2)  #加载程序包
ggplot(role_counts, aes(x = role, y = count, fill = role)) + geom_bar(stat = "identity", width = 0.75) + xlab("人物")+ylab("频数")