Bag-of-words-representation

Introduktion till naturlig språkbehandling i R

Kasey Jones

Research Data Scientist

Föregående exempel

animal_farm %>%
  unnest_tokens(output = "word", token = "words",
                input = text_column) %>%
  anti_join(stop_words) %>%
  count(word, sort = TRUE)

# A tibble: 3,611 x 2
   word         n
   <chr>    <int>
 1 animals    248
 2 farm       163
 ...

Bag-of-words-representation

text1 <- c("Few words are important.")
text2 <- c("All words are important.")
text3 <- c("Most words are important.")

Unika ord:

few: bara i text1
all: bara i text2
most: bara i text3
words, are, important

Typiska vektorrepresentationer

# Lowercase, without stop words
word_vector <- c("few", "all", "most", "words", "important")

# Representation for text1
text1 <- c("Few words are important.")
text1_vector <- c(1, 0, 0, 1, 1)
# Representation for text2
text2 <- c("All words are important.")
text2_vector <- c(0, 1, 0, 1, 1)
# Representation for text3
text3 <- c("Most words are important.")
text3_vector <- c(0, 0, 1, 1, 1)

tidytext-representation

words <- animal_farm %>%
    unnest_tokens(output = "word", token = "words", input = text_column) %>%
    anti_join(stop_words) %>%
    count(chapter, word, sort = TRUE)
words

# A tibble: 6,807 x 3
   chapter    word         n
   <chr>      <chr>    <int>
 1 Chapter 8  napoleon    43
 2 Chapter 8  animals     41
 3 Chapter 9  boxer       34
...

Exempel med ett ord

words %>%
  filter(word == 'napoleon') %>%
  arrange(desc(n))

# A tibble: 9 x 3
  chapter    word         n
  <chr>      <chr>    <int>
1 Chapter 8  napoleon    43
2 Chapter 7  napoleon    24
3 Chapter 5  napoleon    22
...
8 Chapter 3  napoleon     3
9 Chapter 4  napoleon     1

Glesa matriser

library(tidytext); library(dplyr)
russian_tweets <- read.csv("russian_1.csv")
russian_tweets <- as_tibble(russian_tweets)

tidy_tweets <- russian_tweets %>%
  unnest_tokens(word, content) %>%
  anti_join(stop_words)
tidy_tweets %>%
  count(word, sort = TRUE)

# A tibble: 43,666 x 2
...

Glesa matriser – fortsättning

Gles matris

20 000 rader (tweets)
43 000 kolumner (ord)
20 000 * 43 000 = 860 000 000
Endast 177 000 poster är skilda från 0. Cirka 0,02 %

Exempel på gles matris: Glesa matriser har mycket få poster skilda från 0. En gles matris består till övervägande del av nollor.

Nu kör vi en övning!

Introduktion till naturlig språkbehandling i R