R中的文本挖掘

Question

我想從包含特定句子（兩個或多個單詞組合）而不是單個單詞的文本創建TDM。 這些句子可以是"climate change" ， "global worming" ， "lad use"等。我所看到的例子都是單個單詞。

tabela = DocumentTermMatrix(textolimpo,
         list(dictionary = c("climate change","global worming","land use")))

我感謝有人能幫助我。

干杯。

拉斐爾

Answer 1

我推薦quanteda ：

library(quanteda)
textolimpo <- c("This climate change concerns me. This climate changes", "Wormed: global worming increased")
(dfm <- dfm(textolimpo, 
           ngrams=2L,
           dictionary = list(climate="climate_change",
                             warm="global_worming"), 
           valuetype = "regex"))
# 2 x 2 sparse Matrix of class "dfmSparse"
#        features
# docs    climate warm
#   text1       2    0
#   text2       0    1
(dfm <- dfm(textolimpo, 
            ngrams=2L, 
            thesaurus = list(climate="climate_change",
                             warm="global_worming"), 
            valuetype = "regex"))
# 2 x 8 sparse Matrix of class "dfmSparse"
#       this_climate change_concerns concerns_me me_this wormed_global worming_increased CLIMATE WARM
# text1            2               1           1       1             0                 0       2    0
# text2            0               0           0       0             1                 1       0    1