Automated Literature Screening for Systematic Reviews: Dataset for Evaluation Against Human Title and Abstract and Full-Text Screening Decisions

This Zenodo entry contains the supplementary material associated with the manuscript titled Automated Literature Screening for Systematic Reviews: A 5-Tier Prompting Approach Meeting Cochrane’s Sensitivity Requirement of Greater Than 0.99. The paper will be presented at LLMsMI 2024 in November 2024.

A script is provided for replicating the executed experiments, along with a comprehensive evaluation file that reports all the experiment results. Provided data files represent an extension to the original datasets as provided by [1]. For associated systematic review manuscripts and eligibility criteria, please refer to [1] as well. 

[1] Guo, Eddie; Gupta, Mehul; Deng, Jiawen; Park, Ye-Jean; Paget, Mike; Naugler, Christopher (2023). "Automated Paper Screening for Clinical Reviews Using Large Language Models." Mendeley Data, V1, doi: 10.17632/np79tmhkh5.1. Accessed from: https://data.mendeley.com/datasets/np79tmhkh5/1.