Large language model-based paper classification framework with key-insight extraction and confidence-weighted voting
1-score of 0.796, substantially exceeding KBC (0.732), ABC (0.676), and unsupervised K-means clustering (0.446). By employing zero-shot LLMs, our approach demonstrated the potential for enhanced adaptability across diverse domains and classification tasks without requiring fine-tuning, demonstrating that a carefully designed pipeline can enable LLMs to achieve classification performance comparable to human reviewers.
Authors
- Wenyang Li (ORCID: https://orcid.org/0000-0001-6078-6357)
- Zihan Song (ORCID: https://orcid.org/0000-0001-5055-179X)
- Xin Zhang (ORCID: https://orcid.org/0000-0002-6507-2975)
- Byung-Kwon Park
- Ngeemasara Thapa
- Bei Zhan
- Wenbin Liu
- Jianfei Li
- Shan Huang
- Jie Lu
Institutions
- ZheJiang Academy of Agricultural Sciences (CN)
- Trust for Advancement of Agricultural Sciences (IN)
- Hubei Engineering University (CN)
- Dong-A University (KR)
Publication Details
- Journal
- Research Synthesis Methods
- Published
- 2026-04-22
- DOI
- https://doi.org/10.1017/rsm.2026.10094
- Primary Topic
- Meta-analysis and systematic reviews
- Type
- article
- Field-Weighted Citation Impact
- 0.00