The exploration mode can be used to explore to performance of the active learning software and the performance of different algorithms on already labeled data. In this mode relevant records are displayed in green and a recall curve can be obtained.
It is assumed you have already installed Python and ASReview. If this is not the case, see Installation. Also, you should have created a project - the name is not relevant, but is advised to have a explore-prefix.
Upload a Benchmark Dataset¶
Select one of the available benchmark-datasets.
In the next step, you are asked to add prior inclusions. Select 1-5 papers of your choice. For the featured datasets you can use the following titles of papers:
Latent trajectories of trauma symptoms and resilience: the 3-year longitudinal prospective USPER study of Danish veterans deployed in Afghanistan
A Latent Growth Mixture Modeling Approach to PTSD Symptoms in Rape Victims
Peace and War: Trajectories of Posttraumatic Stress Disorder Symptoms Before, During, and After Military Deployment in Afghanistan
The relationship between course of PTSD symptoms in deployed U.S. Marines and degree of combat exposure
Trajectories of trauma symptoms and resilience in deployed US military service members: Prospective cohort study
Detection of viromes of RNA viruses using the next generation sequencing libraries prepared by three methods
Identification of a novel astrovirus in domestic sheep in Hungary
Identification of novel bovine group A rotavirus G15P strain from epizootic diarrhea of adult cows by de novo sequencing using a next-generation sequencer
Detection of novel viruses in porcine fecal samples from China
Metagenomic Analysis of the Jinding Duck Fecal Virome
Software Fault Prediction:
Predicting Defect-Prone Software Modules at Different Logical Levels
Quantitative analysis of faults and failures in a complex software system
A Comprehensive Empirical Study of Count Models for Software Fault Prediction
Predicting fault prone modules by the Dempster-Shafer belief networks
Robust prediction of fault-proneness by random forests
Quinapril in patients with congestive heart failure: controlled trial versus captopril.
Clinical effects of early angiotensin-converting enzyme inhibitor treatment for acute myocardial infarction are similar in the presence and absence of aspirin: systematic overview of individual data from 96,712 randomized patients. Angiotensin-converting Enzyme Inhibitor Myocardial Infarction Collaborative Group.
Efficacy of different drug classes used to initiate antihypertensive treatment in black subjects: results of a randomized trial in Johannesburg, South Africa.
Long-term mortality in patients with myocardial infarction: impact of early treatment with captopril for 4 weeks.
Comparison of perindopril versus captopril for treatment of acute myocardial infarction.
Mark five random papers as irrelevant.
Start reviewing the first 50, 100 or even 200 papers. Abstracts in green are relevenant papers and abstracts in black are irrelevant.
For the PTSD Trajectories dataset you expect to find about 7 out of 38 relevant papers after screening 50 papers, 19 after screening 100 papers and 36 after 200 papers.
For the Virus Metagenomics dataset you expect to find 20 out of 120 relevant papers after screening 50 papers, 40 after screening 100 papers and 70 after 200 papers
For the Software Fault Prediction dataset you expect to find 25 out of 104 relevant papers after screening 50 papers, 48 after screening 100 papers and 88 after 200 papers.
For the ACEinhibitors dataset you expect to find 16 out of 41 relevant papers after screening 50 papers, 27 after screening 100 papers and 32 after 200 papers.
Upload Your own Data for Exploration¶
You can explore a previously labeled dataset in ASReview LAB by adding an extra column called ‘debug_label’ to your dataset. This column should indicate the relevant (1) and irrelevant (0) records. The relevant records will show up green during screening.
Open ASReview LAB.
Start a new project.
Click the Start Setup button.
Select your labeled dataset containing the ‘debug_label’.