ASReview
ASReview
Open-source machine-learning software that uses active learning to prioritize records during systematic-review screening while keeping the researcher in control of labeling decisions.
Developer:
Utrecht University
Our Take
ASReview is an open-source option for prioritizing records during systematic-review screening using active learning. As reviewers label relevant and irrelevant records, the system can help rank the remaining set so that likely-relevant studies are considered earlier in the process.
Ranking is a prioritization aid, not a guarantee that all relevant studies will be found or that unreviewed records can safely be discarded. Results depend on the initial labels, dataset, and screening protocol, so transparency about the workflow is essential.
Define eligibility criteria first, document the software version and settings, and retain the full search set and screening decisions. Follow the stopping and validation approach required by your review method, and report how machine-learning prioritization affected the screening process.
Key Features
Active-learning screening, Iterative relevance ranking, Configurable machine-learning models, Screening progress visualization
Best For
AI-assisted screening of large literature sets
Research Tasks
Reduce manual screening effort in large literature searches; prioritize likely relevant records; test and document screening strategies; verify stopping decisions and missed-study risk.
Pricing
$0 — completely free and open source, with no subscription or licensing fee. It runs locally by default; hosting it on a server may incur infrastructure costs.