In elizagrames/synthesisr: Import, Assemble, and Deduplicate Systematic Review Search Results

knitr::opts_chunk$set(echo = TRUE)

Introduction to synthesisr

Import results and remove duplicates

my_directory <- "~/synthesisr/inst/extdata/"

search_results <- synthesisr::import_results(directory=my_directory, 
                                             filename = NULL, 
                                             save_dataset = FALSE, 
                                             verbose = TRUE)

first_dedupe <- synthesisr::deduplicate(df=search_results, 
                                        field = "title", 
                                        method = "quick", 
                                        language = "English")

second_dedupe <- synthesisr::deduplicate(df=first_dedupe, 
                                         field="title",
                                         method="similarity",
                                         language="English",
                                         cutoff_distance = 2)

Detect chimeras

Before accepting our deduped dataset, we want to check for any multilingual, chimera titles that could be limiting the functionality of the dedupe function. If chimeras are detected, they should be split into single-language titles and passed back to the deduplicate function. An automated function to trim chimera text is pending development for the next version of synthesisr.

any_chimeras <- synthesisr::chimera_detect(second_dedupe$title, overlap = .6)

Create document-feature matrices

my_dfm <- synthesisr::create_dfm(second_dedupe$abstract, language="English")

Report search results

elizagrames/synthesisr documentation built on May 26, 2019, 10:34 a.m.