In kasperwelbers/corpus-tools: tools for corpus analysis

Compare multiple corpora

The corpora.compare function is explained in the compare howto. Here we show how corpora.list.compare can be used to compare many corpora. Also, we introduce the documents.compare and documents.window.compare functions.

This is currently under development.

library(corpustools)
data(sotu)
head(sotu.tokens)
dtm = dtm.create(sotu.tokens$aid, sotu.tokens$lemma)
meta = sotu.meta[match(rownames(dtm), sotu.meta$id),]

### comparing corpora ###
permediumcomparison = corpora.compare.list(dtm, subcorpus=meta$headline)
peryearcomparison = corpora.compare.list(dtm, subcorpus=meta$date, method='window')

# or using a dtm list as input
dtm_list = splitDtm(dtm, subcorpus=meta$date)
names(dtm_list)
peryearcomparison = corpora.compare.list(dtm_list, method='window')

### comparing documents ###
d = documents.compare(dtm, min.similarity=0.8)
head(d)
terms = getOverlapTerms(d$x, d$y, dtm.x=dtm)
head(terms)

# compare in windows (days)
d = documents.window.compare(dtm, meta$date, window.size=2, window.direction='<', time.unit='days')
head(d)