知识卡片
同名多标签不能靠删记录消除
同一姓名可以在两个登记类别出现,表明输入不足以确定唯一标签。取较多数类别只改变任务定义,不消除真实歧义;需说明预测的是多数标签、概率分布还是个体记录,并相应定义损失和基准。
这张卡片看不懂?
这张卡片已通过人工审查,不再接受失败反馈。
已有反馈 · 0 条
还没有人提交反馈。
在整书笔记中的位置
- · 神经网络基础
原书出处 · 1 处
- · 查看原书依据
You want to make sure that your test set names don’t appear anywhere in your training set. You also want to make sure that your test set only has one “right” label for each name. But there is not one correct binary sex label for any particular name; there is, indeed, a probability score (continuous value) of “maleness” or “femaleness” of a name, based on
参考来源
- 位置:《Natural Language Processing in Action, Second Edition》第05章《神经网络基础》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Natural Language Processing in Action, Second Edition/_epub-src/OEBPS/Text/chapter-5.html)
- 结论依据:摘录中关于同名多标签不能靠删记录消除的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。
- 原始内容:You want to make sure that your test set names don’t appear anywhere in your training set. You also want to make sure that your test set only has one “right” label for each name. But there is not one correct binary sex label for any particular name; there is, indeed, a probability score (continuous value) of “maleness” or “femaleness” of a name, based on