quanteda (version 0.9.9-50)

dfm_compress: compress a dfm or fcm by combining identical dimension elements

Description

"Compresses" a dfm or fcm whose dimension names are the same, for either documents or features. This may happen, for instance, if features are made equivalent through application of a thesaurus. It may also occur after lower-casing or stemming the features of a dfm, but this should only be done in very rare cases (approaching never: it's better to do this before constructing the dfm.) It could also be needed after a cbind.dfm or rbind.dfm operation.

Usage

dfm_compress(x, margin = c("both", "documents", "features"))
fcm_compress(x)

Arguments

x

input object, a dfm or fcm

margin

character indicating on which margin to compress a dfm, either "documents", "features", or "both" (default). For fcm objects, "documents" has no effect.

...

additional arguments passed from generic to specific methods

Examples

mat <- rbind(dfm(c("b A A", "C C a b B"), tolower = FALSE, verbose = FALSE),
             dfm("A C C C C C", tolower = FALSE, verbose = FALSE))
colnames(mat) <- char_tolower(featnames(mat))
mat
dfm_compress(mat, margin = "documents")
dfm_compress(mat, margin = "features")
dfm_compress(mat)

# no effect if no compression needed
compactdfm <- dfm(data_corpus_inaugural[1:5])
dim(compactdfm)
dim(dfm_compress(compactdfm))

# compress an fcm
myfcm <- fcm(tokens("A D a C E a d F e B A C E D"), 
             context = "window", window = 3)
## this will produce an error:
# fcm_compress(myfcm)

txt <- c("The fox JUMPED over the dog.",
         "The dog jumped over the fox.")
toks <- tokens(txt, remove_punct = TRUE)
myfcm <- fcm(toks, context = "document")
colnames(myfcm) <- rownames(myfcm) <- tolower(colnames(myfcm))
colnames(myfcm)[5] <- rownames(myfcm)[5] <- "fox"
myfcm
fcm_compress(myfcm)