R TermDocumentMatrix 단어 병합

1 개요

R TermDocumentMatrix 용어 병합

zTdmMergeTerms()
- 1개 또는 여러 개의 기존용어(froms)를 하나의 신규용어(to)로 병합(합산)한다.
- 새로운 용어(to)가 tdm 내에 있어도 되고 없어도 된다.
- 기존용어(froms)에 to가 들어있는 경우 무시된다.
zTdmMergeTermsGrep()
- 정규식으로 잡히는 용어들을 하나로 병합(합산)한다.

library(tm) # TermDocumentMatrix()

zTdmMergeTerms <- function(tdm, froms, to) {
  froms <- froms[froms!=to & froms %in% rownames(tdm)]
  if(length(froms)<1) return(tdm)
  if(!to %in% rownames(tdm)) {
    tdm$nrow <- as.integer(tdm$nrow + 1)
    rownames(tdm) <- append(rownames(tdm),to)
  }
  for(from in froms) {
    nms <- rownames(tdm)
    t1 <- tdm[which(nms==from),] + tdm[which(nms==to),]
    rownames(t1) <- to
    t2 <- tdm[which(nms!=from & nms !=to),]
    t1$i <- t1$i + t2$nrow
    t2$i <- c(t2$i, t1$i)
    t2$j <- c(t2$j, t1$j)
    t2$v <- c(t2$v, t1$v)
    t2$nrow <- as.integer(t2$nrow + 1)
    rownames(t2) <- append(rownames(t2),to)
    tdm <- t2
  }
  return(tdm)
}
zTdmMergeTermsGrep <- function(tdm, pattern, to) {
  return(zTdmMergeTerms(tdm, rownames(tdm)[grep(pattern,rownames(tdm))], to))
}

#### tdm1
docs1 = c("안녕, 안녕.", "반갑다, 친구들.", "반갑다, 친구야, 친구야")
## docs1 = iconv(docs1, "CP949", "UTF-8") ## 윈도우 R스튜디오 하드코딩 데이터입력시 인코딩 변환
tdm1 = TermDocumentMatrix(Corpus(VectorSource(docs1)),control=list(removePunctuation=T,stopwords=F))
inspect(tdm1)
## <<TermDocumentMatrix (terms: 4, documents: 3)>>
## Non-/sparse entries: 5/7
## Sparsity           : 58%
## Maximal term length: 3
## Weighting          : term frequency (tf)
## Sample             :
##         Docs
## Terms    1 2 3
##   반갑다 0 1 1
##   안녕   2 0 0
##   친구들 0 1 0
##   친구야 0 0 2

inspect(zTdmMergeTerms(tdm1,"반갑다","안녕"))
## <<TermDocumentMatrix (terms: 3, documents: 3)>>
## Non-/sparse entries: 5/4
## Sparsity           : 44%
## Maximal term length: 3
## Weighting          : term frequency (tf)
## Sample             :
##         Docs
## Terms    1 2 3
##   안녕   2 1 1
##   친구들 0 1 0
##   친구야 0 0 2

inspect(zTdmMergeTerms(tdm1,c("반갑다","안녕","친구야"),"인사"))
## <<TermDocumentMatrix (terms: 2, documents: 3)>>
## Non-/sparse entries: 4/2
## Sparsity           : 33%
## Maximal term length: 3
## Weighting          : term frequency (tf)
## Sample             :
##         Docs
## Terms    1 2 3
##   인사   2 1 3
##   친구들 0 1 0

inspect(zTdmMergeTerms(tdm1,c("반갑다","안녕","친구야"),"안녕"))
## <<TermDocumentMatrix (terms: 2, documents: 3)>>
## Non-/sparse entries: 4/2
## Sparsity           : 33%
## Maximal term length: 3
## Weighting          : term frequency (tf)
## Sample             :
##         Docs
## Terms    1 2 3
##   안녕   2 1 3
##   친구들 0 1 0

inspect(zTdmMergeTermsGrep(tdm1, "^친구", "친구"))
## <<TermDocumentMatrix (terms: 3, documents: 3)>>
## Non-/sparse entries: 5/4
## Sparsity           : 44%
## Maximal term length: 3
## Weighting          : term frequency (tf)
## Sample             :
##         Docs
## Terms    1 2 3
##   반갑다 0 1 1
##   안녕   2 0 0
##   친구   0 1 2

2 같이 보기