S4 class to organize counts. The classes polmineR and
ngrams inherit from the class.
The summary-method in combination with a weighed
count-object can be used to perform a dictionary-based sentiment
analysis (see examples).
The length-method is synonymous with the size-method
and will return the size of the corpus or partition a count
has been derived from.
statObject of class data.table.
corpusObject of class character the CWB corpus the partition is based on .
encodingObject of class character, the encoding of the corpus.
nameObject of class character, a name for the object.
sizeObject of class integer, the size of the partition or
corpus the count is based upon.
The count-class inherits from the textstat-class.
# sample for dictionary-based sentiment analysis
weights <- data.table::data.table(
word = c("gut", "super", "herrlich", "schlecht", "grob", "mies"),
weight = c(1,1,1,-1,-1,-1)
)
corp <- corpus("GERMAPARLMINI")
sc <- subset(corp, date == "2009-11-11")
cnt <- count(sc, p_attribute = "word")
cnt <- weigh(cnt, with = weights)
y <- summary(cnt)
# old, partition-based workflow
p <- partition("GERMAPARLMINI", date = "2009-11-11")
#> ... get cpos and strucs
p <- enrich(p, p_attribute = "word")
#> ... getting counts for p-attribute(s): word
weights <- data.table::data.table(
word = c("gut", "super", "herrlich", "schlecht", "grob", "mies"),
weight = c(1,1,1,-1,-1,-1)
)
p <- weigh(p, with = weights)
summary(p)
#> $name
#> [1] ""
#>
#> $size
#> [1] 117614
#>
#> $p_attribute
#> [1] "word"
#>
#> $unique
#> [1] 12080
#>
#> $positive_n
#> [1] 87
#>
#> $positive_share
#> [1] 0.0007397079
#>
#> $positive_weighed
#> [1] 87
#>
#> $negative_n
#> [1] 7
#>
#> $negative_share
#> [1] 5.951672e-05
#>
#> $negative_weighed
#> [1] -7
#>