Unverified50% confidenceTradeoffExact time
在医疗诊断中假阴性代价极高因此召回率比精确率更重要，在垃圾邮件过滤中假阳性代价较高精确率更受关注
1
Sources
50%
Confidence
Long-term
Relevance
8/29/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIn the no-show prediction task, the cost of missing truly absent patients (false negatives) is far higher than the cost of misclassifying patients who do show up, so recall should be prioritized.75% similarUnverified在专业信息上对模型进行微调可能导致更多幻觉，且成本效益极差70% similarUnverified对于疟疾诊断场景，假阴性的危害远大于假阳性，因此召回率（Recall/Sensitivity）比整体准确率更为关键69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/816324API
curl https://kongchang.com/api/v1/knowledge/claims/816324MCP
get_claim(id=816324)