step2.R

# how often are words (word-stems) used across all the docs
dtm <- DocumentTermMatrix(corpus.stemmed)

# in the first 5 text files, how frequent are the first 8 words (alphabetical order)
inspect(dtm[1:5, 1:8])

# let's make that dtm table a matrix...
dtm.mat <- as.matrix(dtm)

####### STEP 3 ----- visualizing the high-frequency words

library(wordcloud)
wordcloud(colnames(dtm.mat), dtm.mat[12, ], max.words = 20) #essay no. 12
wordcloud(colnames(dtm.mat), dtm.mat[24, ], max.words = 20) #essay no. 24


# If we forget what a word stem refers to then we can find out
stemCompletion(c("revenu", "commerc",
                 "peac", "army"), corpus.prep)


## Here's a way of figuring out how important a word is, in a particular doc
dtm.tfidf <- weightTfIdf(dtm) # tf-idf calculation (an importance measure)

dtm.tfidf.mat <- as.matrix(dtm.tfidf) #convert to matrix

## 10 most important words for paper no. 12
head(sort(dtm.tfidf.mat[12, ], decreasing = TRUE), n = 10)

## 10 most important words for paper no. 24
head(sort(dtm.tfidf.mat[24, ], decreasing = TRUE), n = 10)