strsim

R package with string similarity functions relevant for psycholinguistic research (e.g. neighborhood density, neighborhood frequency, OLD20).

Installation

library(devtools)
devtools::install_github("waltervanheuven/strsim")

Usage

library(strsim)
library(readxl)

# download SUBTLEX-UK database (van Heuven et al., 2014) that contains a large list of English words
download.file(url='https://psychology.nottingham.ac.uk/subtlex-uk/SUBTLEX-UK.xlsx.zip',
              destfile='SUBTLEX-UK.xlsx.zip',
              method='curl')
# unzip file              
unzip("SUBTLEX-UK.xlsx.zip")
# load xlsx
subtlex_uk <- read_xlsx("SUBTLEX-UK.xlsx")

# Orthographic neighbors of the word book
neighborhood_density("book", subtlex_uk$Spelling, show=TRUE)

# Orthographic neighbors of all words
# progress bar is shown and multiple cores are used to speed up calculation
nd <- neighborhood_density(subtlex_uk$Spelling, subtlex_uk$Spelling, pb=TRUE, parallel=TRUE)

# Neighborhood frequency
neighborhood_frequency("book", subtlex_uk$Spelling, subtlex_uk$`LogFreq(Zipf)`, show=TRUE)

# OLD20 value and the 20 orthographically closest words of the word book
# based on the Levenshtein distance
old("book", subtlex_uk$Spelling, show=TRUE)

# calculate OLD20 for a list of words
old20(c("room", "apricot", "dictionary"), subtlex_uk$Spelling)

# calculate_old20 function utilises multiple cores to improve performance
# when calculating old20 for large number of words, e.g.
old20 <- old20(subtlex_uk$Spelling, subtlex_uk$Spelling, pb=TRUE, parallel=TRUE)