eztfidf: Turn character vectors into tfidf distances easily.

In patricklyngrutz/eztfidf: Simple interface for TF-IDF distances

Description

This function returns an eztfidf list containing convenient functions.

Usage

eztfidf(char_vector, replace_words = c(`\t` = " "))

Arguments

char_vector	A character vector of documents. To be passed as a VectorSource (tm package). The values may be duplicated but the names may not.
replace_words	A named character vector. The element names will be replaced with the elements.

Examples

super_heroes <- c(
    'The Flash', 'The HULK', 'she-hulk', 'ant-man', 'Ironman', 'BATMAN',
    'superman', 'the green arrow', 'aqua-man', 'the silver surfer', 'green lantern'
    )
names(super_heroes) <- super_heroes
super_heroes <- gsub('man$', '-MAN', super_heroes, TRUE)  # custom cleaning
x <- eztfidf(
    super_heroes, replace_words = c('-' = ' ', 'silver' = 'gold')
)

# Use numeric index or original names to see changes to docs
x$docs[1:10]
x$docs[c('The HULK','the silver surfer')]

# Inspect bag-of-words tfidf values as a list or matrix
x$values(c('the green arrow','green lantern'))
x$values(c(2,3,8,11), mode = 'matrix')

# Best matching values and cosine similarity matrix easily accessible
x$CosineSimVector(3, top = 3)
x$CosineSimVector('the green arrow', top = 3)
x$CosineSimMatrix(c(2,3,8,11))

patricklyngrutz/eztfidf documentation built on May 6, 2019, 8:31 p.m.