dist.cosine: Cosine Distance

In computationalstylistics/stylo: Stylometric Multivariate Analyses

Cosine Distance

Description

Function for computing a cosine similarity of a matrix of values, e.g. a table of word frequencies. Recent findings (Jannidis et al. 2015) show that this distance outperforms other nearest neighbor approaches in the domain of authorship attribution.

Usage

dist.cosine(x)

Arguments

x	a matrix or data table containing at least 2 rows and 2 cols, the samples (texts) to be compared in rows, the variables in columns.

Value

The function returns an object of the class dist, containing distances between each pair of samples. To convert it to a square matrix instead, use the generic function as.dist.

Author(s)

Maciej Eder

References

Evert, S., Proisl, T., Jannidis, F., Reger, I., Pielstrom, S., Schoch, C. and Vitt, T. (2017). Understanding and explaining Delta measures for authorship attribution. Digital Scholarship in the Humanities, 32(suppl. 2): 4-16.

See Also

stylo, classify, dist, as.dist

Examples

# first, preparing a table of word frequencies
        Iuvenalis_1 = c(3.939, 0.635, 1.143, 0.762, 0.423)
        Iuvenalis_2 = c(3.733, 0.822, 1.066, 0.933, 0.511)
        Tibullus_1  = c(2.835, 1.302, 0.804, 0.862, 0.881)
        Tibullus_2  = c(2.911, 0.436, 0.400, 0.946, 0.618)
        Tibullus_3  = c(1.893, 1.082, 0.991, 0.879, 1.487)
        dataset = rbind(Iuvenalis_1, Iuvenalis_2, Tibullus_1, Tibullus_2, 
                        Tibullus_3)
        colnames(dataset) = c("et", "non", "in", "est", "nec")

# the table of frequencies looks as follows
        print(dataset)
        
# then, applying a distance, in two flavors
        dist.cosine(dataset)
        as.matrix(dist.cosine(dataset))

computationalstylistics/stylo documentation built on April 7, 2024, 4:12 p.m.