PubChat Benchmark Validation Data, Protocols, and Reproducibility Materials

Le résumé fourni par la source

This record contains the data tables, protocol documentation, prompt templates, study instruments, and reproducibility materials supporting the manuscript “Development and Benchmark Validation of PubChat for PubMed-Grounded Multilingual Biomedical LiteratureRetrieval”. The archive is organized into seven numbered folders: 1. PubChat 8-Language Retrieval Output for 20 CDSR Reviews: Benchmark definitions and per-CDSR multilingual PubChat retrieval outputs. 2. PubChat vs 8-Tool Benchmarking Data with Evaluation Protocols: Raw benchmarking data and tool-evaluation protocol documentation. 3. PubChat Prompt Templates and Reproducibility Details: Prompt templates, runtime placeholder definitions, and reproducibility documentation. 4. Post Hoc Statistical Analysis of Tool Benchmarking Results: Friedman omnibus and Wilcoxon-Holm post hoc statistical analysis tables. 5. PubChat Retrieval Stress-Testing and Boundary Analyses: Stress-testing outputs, boundary-analysis tables, and protocol documentation. 6. PubChat User Survey - 42-Item Questionnaire: User-assessment questionnaire and assessment protocol. 7. MIDE Case Study - Periodontitis and CKD: MIDE case-study corpus, configuration, novelty scores, and generated mini-review output. Protocol documents and prompt templates are provided for transparency and reproducibility. Core methodological details are reported in the manuscript Methods section. Source code repositories:PubChat: https://github.com/PubChatOfficial/PubChat_smart_literature_searchMIDE: https://github.com/PubChatOfficial/MIDE-skill

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.