Lecture7: Network representation, TwitteR
文字単位のn-gram:英文
test1<-"Osaka University said Wednesday that personal data of around 80,000 students"
substr(test1,1,4)## [1] "Osak"size=4
len<-nchar(test1)-size+1
ngramLst <- c()
for(i in 1:nchar(test1)){
ngramLst<-rbind(ngramLst,(substr(test1,i,i+size-1)))
}
head(ngramLst)## [,1]
## [1,] "Osak"
## [2,] "saka"
## [3,] "aka "
## [4,] "ka U"
## [5,] "a Un"
## [6,] " Uni"文字単位のn-gram:日本語
準備
test2<-"大阪大学は１３日、学内システムに海外から不正アクセスがあり、"
substr(test2,1,4)## [1] "大阪大学"nchar(test2)## [1] 30文字単位のngram
size=4
len<-nchar(test2)-size+1
ngramLst <- c()
for(i in 1:len){
ngramLst<-rbind(ngramLst,(substr(test2,i,i+size-1)))
}
head(ngramLst)## [,1]
## [1,] "大阪大学"
## [2,] "阪大学は"
## [3,] "大学は１"
## [4,] "学は１３"
## [5,] "は１３日"
## [6,] "１３日、"単語単位のn-gram
準備：単語単位のリスト作成
wordLst <- strsplit(test1, "[[:space:]]|[[:punct:]]")
wordLst <- unlist(wordLst)
wordLst <- tolower(wordLst)
wordLst <- wordLst[wordLst != ""]
wordLst## [1] "osaka" "university" "said" "wednesday" "that"
## [6] "personal" "data" "of" "around" "80"
## [11] "000" "students"単語単位のngram
size=3
start=1
wordLst[start:(start+size-1)]## [1] "osaka" "university" "said"strLst<-c()
len<-length(wordLst)-size+1
for(i in seq(1,len,size)) {
strLst<-rbind(strLst,wordLst[i:(i+size-1)])
}
strLst## [,1] [,2] [,3]
## [1,] "osaka" "university" "said"
## [2,] "wednesday" "that" "personal"
## [3,] "data" "of" "around"
## [4,] "80" "000" "students"RMecabによる日本語形態素解析
library("RMeCab")
wordLst<-unlist(RMeCabC(test2))
wordLst## 名詞 助詞 名詞 名詞 名詞 記号
## "大阪大学" "は" "１" "３" "日" "、"
## 名詞 名詞 助詞 名詞 助詞 名詞
## "学内" "システム" "に" "海外" "から" "不正"
## 名詞 助詞 動詞 記号
## "アクセス" "が" "あり" "、"記号を除除
wordLst <- wordLst[names(wordLst)!="記号"]
wordLst## 名詞 助詞 名詞 名詞 名詞 名詞
## "大阪大学" "は" "１" "３" "日" "学内"
## 名詞 助詞 名詞 助詞 名詞 名詞
## "システム" "に" "海外" "から" "不正" "アクセス"
## 助詞 動詞
## "が" "あり"単語単位のngram
size=3
len<-length(wordLst)-size+1
strLst<-c()
for(i in seq(1,len,size)) {
strLst<-rbind(strLst,wordLst[i:(i+size-1)])
}
strLst## 名詞 助詞 名詞
## [1,] "大阪大学" "は" "１"
## [2,] "３" "日" "学内"
## [3,] "システム" "に" "海外"
## [4,] "から" "不正" "アクセス"ネットワーク描画
使用データ
wordLst <- strsplit(test1, "[[:space:]]|[[:punct:]]")
wordLst <- unlist(wordLst)
wordLst <- tolower(wordLst)
wordLst <- wordLst[wordLst != ""]
wordLst## [1] "osaka" "university" "said" "wednesday" "that"
## [6] "personal" "data" "of" "around" "80"
## [11] "000" "students"ngramによる区分: ngram=3, step=2
size=3
step=2
strLst<-c()
len<-length(wordLst)-size+1
for(i in seq(1,len,step)) {
strLst<-rbind(strLst,wordLst[i:(i+size-1)])
}
strLst## [,1] [,2] [,3]
## [1,] "osaka" "university" "said"
## [2,] "said" "wednesday" "that"
## [3,] "that" "personal" "data"
## [4,] "data" "of" "around"
## [5,] "around" "80" "000"ペアデータの作成
一行のデータ
str<-strLst[1,]
pLst <- c()
for(i in 1:(length(str)-1)){
for(j in (i+1):length(str)){
tmp<-cbind(str[i],str[j])
pLst<-rbind(pLst,tmp)
}
}
pLst## [,1] [,2]
## [1,] "osaka" "university"
## [2,] "osaka" "said"
## [3,] "university" "said"netwkPairs.R（ペア単語取得用関数）のロード
source("netwkPairs.R")ペアデータの取得
getPairs(strLst[1,])## [,1] [,2]
## [1,] "osaka" "university"
## [2,] "osaka" "said"
## [3,] "university" "said"getPairs(strLst[2,])## [,1] [,2]
## [1,] "said" "wednesday"
## [2,] "said" "that"
## [3,] "wednesday" "that"getPairsLst(strLst)## [,1] [,2]
## [1,] "osaka" "university"
## [2,] "osaka" "said"
## [3,] "university" "said"
## [4,] "said" "wednesday"
## [5,] "said" "that"
## [6,] "wednesday" "that"
## [7,] "that" "personal"
## [8,] "that" "data"
## [9,] "personal" "data"
## [10,] "data" "of"
## [11,] "data" "around"
## [12,] "of" "around"
## [13,] "around" "80"
## [14,] "around" "000"
## [15,] "80" "000"ペア頻度表
pFreq<-getPairsFreq(strLst)
head(pFreq)## Term1 Term2 Freq
## 1 osaka university 1
## 2 osaka said 1
## 3 university said 1
## 4 said wednesday 1
## 5 said that 1
## 6 wednesday that 1igraphをインストールする
install.packages("igraph")igraphを利用した描画（有向グラフ）
library(igraph)##
## Attaching package: 'igraph'## The following objects are masked from 'package:stats':
##
## decompose, spectrum## The following object is masked from 'package:base':
##
## unionwng<-graph.data.frame(pFreq)
plot(wng)
igraphを利用した描画（無向グラフ）
wng<-as.undirected(graph.data.frame(pFreq))
plot(wng)
igraphを利用した描画（エッジ幅とノードの大きさを調整）
wng<-as.undirected(graph.data.frame(pFreq))
E(wng)$weight<-pFreq$Freq
deg<-degree(wng)
plot(wng,edge.width=E(wng)$weight,vertex.size=30*(deg/max(deg)))
ノードの色, フォントサイズ
wng<-as.undirected(graph.data.frame(pFreq))
E(wng)$weight<-pFreq$Freq
deg<-degree(wng)
V(wng)$color <- "lightblue"
V(wng)$shape <- "rectangle"
V(wng)$label.cex <- 1.5
V(wng)$label.color <- "red"
plot(wng,edge.width=E(wng)$weight,vertex.size=30*(deg/max(deg)))
tkplot
tkplot(wng,edge.width=E(wng)$weight,vertex.size=30*(deg/max(deg)))Twitterアプリケーションの作成＆登録
OAuth 認証用
ROAuth, twitteRのインストール
install.packages('twitteR')
install.packages('ROAuth')ROAuth, twitteRのインストールの読み込み
library(twitteR)
library(ROAuth)twitteRからのOauth認証
cacert.pemをダウンロード
download.file(url="http://curl.haxx.se/ca/cacert.pem", destfile="cacert.pem")認証情報(Twitterアプリケーション)
consr_key="***********"
consr_secrt="***********"
req_url ="https://api.twitter.com/oauth/request_token"
acs_url = "https://api.twitter.com/oauth/access_token"
auth_url="https://api.twitter.com/oauth/authorize"cred<-OAuthFactory$new(consumerKey=consr_key,consumerSecret=consr_secrt,requestURL =req_url,accessURL = acs_url,authURL=auth_url)handshake: twitterクライアント接続
cred$handshake(cainfo="cacert.pem")認証情報取得
setup_twitter_oauth(consr_key, consr_secrt, acs_token, acs_token_sec)検索例
searchTwitter("#DH2018", n=10)
searchTwitter("#永世七冠", n=10)
username = "casualconc"
userTimeline(username)
username = "langstat"
userTimeline(username)