#classification

• 先判定纠错类型再谈规则 20260424232347000
• 二层规则先看强度再看重叠 20260424232219000
• 综述只负责引路 20260424205350000
• PMID 33977284 is RTI incidence endpoint, not direct longevity evidence 20260424204839000
• PMID 25540326 is immune-response endpoint, not longevity evidence 20260424204718000
• 综述只能当入口，不能当证据 20260424204557000
• 先锁原始试验再谈长寿终点 20260424203542000
• PEARL trial is not direct longevity evidence 20260424203525000
• 先用终点语境判定是否真是长寿证据 20260424203253000
• Iter-4360dd15-0078-two-pmid-evidence-verification 20260424203242000
• Iter-4360dd15-0075-abstract-field-extraction-two-pmids 20260424202951000
• 先抽字段再分类 20260424202727000
• 先做摘要级字段抽取 20260424202425000
• Iter-4360dd15-0061-candidate-audit-10-remaining-pmids 20260424200221000
• 先拿原始候选表再分类 20260424200123000
• PMID 38310895 is a review, not a primary trial 20260424200109000
• 边界术语先做例外判定 20260424195952000
• PMID 35040506 longitudinal analysis is embedded in primary trial 20260424195939000
• 先分类型再写表 20260424195838000
• Iter-4360dd15-0058-longevity-evidence-proximity-tagging 20260424195827000
• Iter-4360dd15-0057-study-design-taxonomy-for-retained-trials 20260424195711000