bibliometrixExtra: R for synonyms
:::success
Build a customed package bibliometrixExtra (Extended functions for bibliometrix) to put the functions in, that will make the synonyms replacing processing easier.
:::
Table of Contents
:::spoiler older info
Need to replace synonyms, but the function termExtraction() in R-bibliometrx can’t work smoothly??
★ replace synonyms by importing csv file syn_replace()
(copy/paste and run this func.)
the customed str_replace_multiple can use multiple patterns to be replaced, but it’s not enough. we expect a replacing function can work like in VOSviewr. the syn_replace can do it.
==version 3, it can replace the full terms, and a term inculeds ‘-’ or ' ‘(space).==
Example:
if we want to replace ‘citaion’ as ‘apple’, then
- ‘cocitaion’ –> ‘cocitaion’
- ‘co-citaion’ –> ‘co-citaion’
- ‘co citation’ –> ‘co citation’
- ‘ciation’ –> ‘apple’
## version 3
syn_replace <- function(df, data, tag){
data[[tag]] <- gsub(
';_', replacement = '; ',
x= gsub('^_', replacement = ' ',
x= gsub('\\s', replacement = '_',
x= gsub('-', replacement = '_-_', data[[tag]]
) ) ) )
for (i in 1:nrow(df)){
data[[tag]] <- gsub(
x = data[[tag]] ,
pattern = paste0( '\\<', df[i, 'pattern'], '\\>' ),
replacement = df[i, 'replace'])
}
data[[tag]] <- gsub('_-_', replacement = '-', x=data[[tag]] )
data[[tag]] <- gsub('_', replacement = ' ', x=data[[tag]] )
return( data[[tag]] )
}
:::
★ Usage
remotes::install_github('tsai-jiewen/bibliometrixExtra')
library(bibliometrixExtra) # syn_export(), syn_replace()
library(bibliometrix)
library(tidyverse)
1. Using syn_export() to export csv from bibliometrix.
(from the customed package bibliometrixExtra )
data(scientometrics, package = "bibliometrixData")
# make a tab-freq table, than export as a csv file
syn_export(
file = 'test1124.csv', # export file name
data = scientometrics, # the biblio data
tag = 'ID' # the field tag
)
the export file looks like that:
:::spoiler older info
-
delete col A (number)
-
change col B name to ‘pattern’ (MUST!)
-
change col C name to ‘replace’ (MUST!), than delete all numbers. :::
-
==fill in the terms you want to replace by in col C (replace), one by one.==
-
==delete the terms no need to change in col B (pattern).==
like this, (for example)
change ‘CITATIONS’ and ‘CITATION ANALYSIS’ to ‘APPLE’.
2. Using syn_import() to read the edited csv file into R env.
DTF <- syn_import(file = 'test1124.csv')
check
> DTF
pattern replace
1 CITATIONS APPLE
2 CITATION ANALYSIS APPLE
3. Using syn_replace() to replace the original synonyms terms.
(from the customed package bibliometrixExtra )
take ‘ID’ for example.
you can change to ‘DE’, etc, but remember keep the $ID and tag='ID' as the same.
scientometrics$ID <- syn_replace(
df = DTF, # the edited file you import
data = scientometrics, # the original biblio data
tag = 'ID' # the field tag
)
:::spoiler older info check before, there were 8 ‘CITATIONS’, 25 ‘CITATION ANALYSIS’ and 0 ‘APPLE’ in the data frame.
> data(scientometrics, package = "bibliometrixData")
> scientometrics$ID %>%
+ stringr::str_match(pattern ='CITATIONS') %>%
+ table()
.
CITATIONS
8
> scientometrics$ID %>%
+ stringr::str_match(pattern = 'CITATION ANALYSIS') %>%
+ table()
.
CITATION ANALYSIS
25
> scientometrics$ID %>%
+ stringr::str_match(pattern = 'APPLE') %>%
+ table()
< table of extent 0 >
check after, there are zero ‘CITATIONS’ and ‘CITATION ANALYSIS’, but 33 ‘APPLE’ now! It’s successful in replacing!
> scientometrics$ID %>%
+ stringr::str_match(pattern = 'CITATIONS') %>%
+ table()
< table of extent 0 >
> scientometrics$ID %>%
+ stringr::str_match(pattern = 'CITATION ANALYSIS') %>%
+ table()
< table of extent 0 >
> scientometrics$ID %>%
+ stringr::str_match(pattern = 'APPLE') %>%
+ table()
.
APPLE
33
:::
Remark
the term-replacing takes place at the bibliometrix data.frame, so it can continue to be used in the next analysis.
thematicEvolution()
Perform a Thematic Evolution Analysis.
- M: can replace as your own biblio dataset.
- years: can set the timepoints as you want
- See thematicEvolution: Perform a Thematic Evolution Analysis
nexus <- thematicEvolution(M,field="DE", years=c(2005, 2010, 2015), n=250, minFreq=2)
plotThematicEvolution(nexus$Nodes,nexus$Edges)
fieldByYear()
Field Tag distribution by Year
fieldByYear(scientometrics, field = "ID", timespan = c(2005,2015),
min.freq = 5, n.items = 5, graph = TRUE)
:::spoiler older info (functions)
(older versions) str_replace_multiple
just a liitle bit better than stringr::str_replace_all
str_replace_multiple <- function(pattern_list, data, tag, replacement){
require(stringr)
for (i in 1:length(pattern_list)){
data[[tag]] <- stringr::str_replace_all(
data[[tag]] ,
pattern = pattern_list[i],
replacement = replacement)
}
return(data[[tag]])
}
(older versions) str_replace_df
## version 2
str_replace_df <- function(df, data, tag){
for (i in 1:nrow(df)){
data[[tag]] <- gsub(
x = data[[tag]] ,
pattern = paste0( '\\<', df[i, 'pattern'], '\\>' ),
replacement = df[i, 'replace'])
}
return(data[[tag]])
}
## version 1
str_replace_df <- function(df, data, tag){
require(stringr)
for (i in 1:nrow(df)){
data[[tag]] <- stringr::str_replace_all(
data[[tag]] ,
pattern = df[i, 'pattern'],
replacement = df[i, 'replace'])
}
return(data[[tag]])
}
:::
