计算 R 中的特定词频

Question

I have a data set where I have split text from the journal abstracts to print 1 word per column. This has lead to over 5 million rows, but I just want certain the word counts of certain words. Below is an example of the data:

• 1 rna
• 1 synthesis
• 1 resembles
• 1 copy
• 1 choice
• 1 rna
• 1 recombination
• 1 process
• 1 nascent
• 1 rna

So in that example let's say I want just the rna counts, I would get 3 and that's it. I have done that word count on the whole set but this is not as useful to me.

wordCount <- m3 %>% count(word, sort = TRUE)

Since many of the words aren't helpful for what I am trying to get to.

Any help would be welcome.

Answer 1

You can group_by the word and count occurrences of each unique word and then subset the ones you want.

library(tidyverse)
data <- data.frame(word = c("rna",
                "synthesis",
                "resembles",
               "copy",
                "choice",
                "rna",
                "recombination",
               "process",
                "nascent",
                "rna"))

counts <- data %>% 
  group_by(word) %>% 
  count()

counts[which(counts$word == "rna"),]

   # A tibble: 1 x 2
# Groups:   word [1]
  word      n
  <fct> <int>
1 rna       3

or using dplyr subsetting:

 counts %>% filter(word == "rna")
# A tibble: 1 x 2
# Groups:   word [1]
  word      n
  <fct> <int>
1 rna       3

Piping it all through at once:

 data %>% 
   group_by(word) %>% 
   count() %>%
   filter(word == "rna")

A one liner with data.table solution:

library(data.table)
setDT(data)
data[word == "rna", .N, by = word]

   word N
1:  rna 3