inst/doc/tf_idf.R

In tidytext: Text Mining using 'dplyr', 'ggplot2', and Other Tidy Tools

## -----------------------------------------------------------------------------
library(knitr)
opts_chunk$set(
  warning = FALSE, message = FALSE,
  eval = requireNamespace("ggplot2", quietly = TRUE)             
)

## -----------------------------------------------------------------------------
library(ggplot2)
theme_set(theme_light())

## -----------------------------------------------------------------------------
library(dplyr)
library(janeaustenr)
library(tidytext)
book_words <- austen_books() %>%
  unnest_tokens(word, text) %>%
  count(book, word, sort = TRUE)

total_words <- book_words %>% group_by(book) %>% summarize(total = sum(n))
book_words <- left_join(book_words, total_words)
book_words

## -----------------------------------------------------------------------------
library(ggplot2)
ggplot(book_words, aes(n/total, fill = book)) +
  geom_histogram(show.legend = FALSE) +
  scale_x_continuous(limits = c(NA, 0.0009)) +
  facet_wrap(vars(book), ncol = 2, scales = "free_y")

## -----------------------------------------------------------------------------
book_words <- book_words %>%
  bind_tf_idf(word, book, n)
book_words

## -----------------------------------------------------------------------------
book_words %>%
  select(-total) %>%
  arrange(desc(tf_idf))

## -----------------------------------------------------------------------------
book_words %>%
  filter(book == "Pride & Prejudice") %>%
  select(-total) %>%
  arrange(desc(tf_idf))