PubMed Cache Data for SR-LLM Query Pipeline Benchmark

Zenodo (CERN European Organization for Nuclear Research)

Abstract

PubMed API response cache used in the benchmarking study: "A Two-Stage LLM Pipeline Using Hypothesis-Accumulated Query Rules for Systematic Review Literature Search: A Multi-Domain Benchmarking Study." This dataset contains cached PubMed search results for 24 held-out systematic reviews across five medical domains (rheumatoid arthritis, oncology, cardiovascular disease, psychiatry, infectious diseases), used in Stage 2 exclusion screening and MedCPT semantic re-ranking. Source code and full reproducibility package: https://github.com/honda-s691470/sr-llm-query-pipeline

Authors 3

1. Suguru Honda
2. Rachel Knevel Aachen

   Newcastle University · Leiden University Medical Center

   Affiliation as printed

   Department of Rheumatology, Leiden University Medical Center, Leiden, Netherlands; Rheumatology, Newcastle University Translational and Clinical Research Institute, Newcastle upon Tyne, UK

3. Erik B. van den Akker Aachen

   Leiden University Medical Center · Delft University of Technology

   Affiliation as printed

   Leiden Computational Biology Centre, Leiden University Medical Center, Leiden, Zuid-Holland, Netherlands; Delft Bioinformatics Lab, Delft University of Technology, Delft, Zuid-Holland, Netherlands