Сравните набор слов в двух документах и ​​найдите совпадающее слово и его частоту во втором документе.

Я беру текст, например, корпуса из интеллектуального анализа текста. Использование пакета tm и функций findFreqTerms,agrep является основным моментом в этом подходе.

agrep

Ищет приблизительные совпадения с шаблоном (первый аргумент) в каждом элементе строки x (второй аргумент), используя обобщенное расстояние редактирования Левенштейна (минимальное возможно взвешенное количество вставок, удалений и замен, необходимых для преобразования одной строки в другую).

Этапы подхода:

тексты -> корпуса -> очистка данных -> findfreqterms -> сравнить с другой матрицей документа термина

library(tm)

c1 <- Corpus(VectorSource("Text mining, also referred to as text data mining, roughly equivalent to text analytics, is the process of deriving high-quality information from text. High-quality information is typically derived through the devising of patterns and trends through means such as statistical pattern learning"))

c2 <- Corpus(VectorSource("Text mining usually involves the process of structuring the input text (usually parsing, along with the addition of some derived linguistic features and the removal of others, and subsequent insertion into a database), deriving patterns within the structured data, and finally evaluation and interpretation of the output"))

c3 <- Corpus(VectorSource("Typical text mining tasks include text categorization, text clustering, concept/entity extraction, production of granular taxonomies, sentiment analysis, document summarization, and entity relation modeling (i.e., learning relations between named entities)"))

# Data Cleaning and transformation
c1 <- tm_map(c1, content_transformer(tolower))
c2 <- tm_map(c2, content_transformer(tolower))
c3 <- tm_map(c3, content_transformer(tolower))

c1 <- tm_map(c1, removePunctuation)
c1 <- tm_map(c1, removeNumbers)
c1 <- tm_map(c1, removeWords, stopwords("english"))
c1 <- tm_map(c1, stripWhitespace)

c2 <- tm_map(c2, removePunctuation)
c2 <- tm_map(c2, removeNumbers)
c2 <- tm_map(c2, removeWords, stopwords("english"))
c2 <- tm_map(c2, stripWhitespace)

c3 <- tm_map(c3, removePunctuation)
c3 <- tm_map(c3, removeNumbers)
c3 <- tm_map(c3, removeWords, stopwords("english"))
c3 <- tm_map(c3, stripWhitespace)

dtm1 <- DocumentTermMatrix(c1, control = list(weighting = weightTfIdf, stopwords = TRUE))
dtm2 <- DocumentTermMatrix(c2, control = list(weighting = weightTfIdf, stopwords = TRUE))
dtm3 <- DocumentTermMatrix(c3, control = list(weighting = weightTfIdf, stopwords = TRUE))

ft1 <- findFreqTerms(dtm1)
ft2 <- findFreqTerms(dtm2)
ft3 <- findFreqTerms(dtm3)

#similarity between c1 and c2
common.c1c2 <- data.frame(term = character(0), freq = integer(0))
for(t in ft1){
  find <- agrep(t, ft2)
  if(length(find) != 0){
    common.c1c2 <- rbind(common.c1c2, data.frame(term = t, freq = length(find)))
  }
}
# Note : this for loop can be substituted by apply family functions if taking time for large text

common.c1c2 содержит общие слова между корпусами 1 и 2 с частотой

> common.c1c2
      term freq
1     also    1
2     data    2
3  derived    1
4 deriving    1
5   mining    1
6  pattern    1
7 patterns    1
8  process    1
9     text    1

> ft1
 [1] "also"        "analytics"   "data"        "derived"     "deriving"    "devising"    "equivalent" 
 [8] "highquality" "information" "learning"    "means"       "mining"      "pattern"     "patterns"   
[15] "process"     "referred"    "roughly"     "statistical" "text"        "trends"      "typically"  

> ft2
 [1] "addition"       "along"          "data"           "database"       "derived"        "deriving"      
 [7] "evaluation"     "features"       "finally"        "input"          "insertion"      "interpretation"
[13] "involves"       "linguistic"     "mining"         "others"         "output"         "parsing"       
[19] "patterns"       "process"        "removal"        "structured"     "structuring"    "subsequent"    
[25] "text"           "usually"        "within"        

Это решение не самое эффективное, но надеюсь, что оно поможет.