Automated Systematic Review Datasets

This project contains datasets for the Automated Systematic Review project. This repository is used to collect, preprocess and share datasets on Systematic Review.

Datasets

The datasets are alphabetically ordered.

Reference	Topic	Sample Size	Inclusion	Link	License
Cohen et al., 2006	ACEInhibitors	2544	1.61%	source	NA
Cohen et al., 2006	ADHD	851	2.35%	source	NA
Cohen et al., 2006	Antihistamines	310	5.16%	source	NA
Cohen et al., 2006	Atypical Antipsychotics	1120	13.04%	source	NA
Cohen et al., 2006	Beta Blockers	2072	2.03%	source	NA
Cohen et al., 2006	Calcium Channel Blockers	1218	8.21%	source	NA
Cohen et al., 2006	Estrogens	368	21.74%	source	NA
Cohen et al., 2006	NSAIDS	393	10.43%	source	NA
Cohen et al., 2006	Opiods	1915	0.78%	source	NA
Cohen et al., 2006	Oral Hypoglycemics	503	27.04%	source	NA
Cohen et al., 2006	Proton Pump Inhibitors	1333	3.83%	source	NA
Cohen et al., 2006	Skeletal Muscle Relaxants	1643	0.55%	source	NA
Cohen et al., 2006	Statins	3465	2.45%	source	NA
Cohen et al., 2006	Triptans	671	3.58%	source	NA
Cohen et al., 2006	Urinary Incontinence	327	12.23%	source	NA
Van de Schoot et al., 2018	PTSD	5783	0.66%	source
Wahono, 2015	Software Defect Detection	7002	0.89%	source	Creative Commons Attribution 4.0 International
Hall et al., 2012	Software Fault Prediction	8911	1.17%	source	Creative Commons Attribution 4.0 International
Radjenović et al., 2013	Software Fault Prediction	6000	0.80%	source	Creative Commons Attribution 4.0 International
Kitchenham et al., 2010	Software Engineering	1704	2.58%	source	Creative Commons Attribution 4.0 International
Bannach-Brown et al., 2019	Animal Model of Depression	1993	14.0%	source	Creative Commons Attribution 4.0 International

How it works

Collecting and preprocessing data

The folder datasets/ has a subfolder for the different Systematic Reviews datasets. Each of these subfolders are little project. They contain code and a README.md. The scripts in the different dataset folder create a subfolder named output/ with the result of the data collection.

Dataset formats

The [Automated Systematic Review](https://github.com/msdslab/automated- systematic-review) software accepts several file formats like RIS and CSV. The datasets in this project are stored in one of these formats.

RIS files

RIS files are used by digital libraries, like IEEE Xplore, Scopus and ScienceDirect. Citation managers Mendeley and EndNote support the RIS format as well. For simulation, we use an additional RIS tag with the letters LI (Label included).

CSV files

For CSV files, the software accepts a set of predetermined labels in line with the ones used in RIS files. The most commonly used ones are: "id", "authors", "date", "title", "keywords" and "abstract". To indicate labelling decisions, one can use "included" or "label_included".

In general, the following column names are recognized (based on https://pypi.org/project/RISparser/):

first_authors
secondary_authors
tertiary_authors
subsidiary_authors
abstract
author_address
accession_number
authors
custom1
custom2
custom3
custom4
custom5
custom6
custom7
custom8
caption
call_number
place_published
date
name_of_database
doi
database_provider
end_page
end_of_reference
edition
id
number
alternate_title1
alternate_title2
alternate_title3
journal_name
keywords
file_attachments1
file_attachments2
figure
language
label
note
type_of_work
notes
abstract
number_of_Volumes
original_publication
publisher
year
reviewed_item
research_notes
reprint_edition
version
issn
start_page
short_title
primary_title
secondary_title
tertiary_title
translated_author
title
translated_title
type_of_reference
unknown_tag
url
volume
publication_year
access_date

The custom tag is:

label_included

Contact and contributors

Contact details can be found at the Automated Systematic Review project page.