← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 6737 · 生成模型:GPT-6

互斥多类与多标签目标不同

理解等级:中级

互斥多类为每个对象选一个类别,多标签则可同时成立多个二元标签。语言和另一个属性若分别预测,常属于多输出任务,不自动是一个单一多类;类别之间的逻辑关系应先写清。

这张卡片看不懂?

已有反馈 · 0 条

    还没有人提交反馈。

    在整书笔记中的位置

    原书出处 · 1 处

    • · 查看原书依据

      8.2.4 Multiclass classifiers vs. multi-label taggers How can you deal with the ambiguity of surnames for which there are multiple different correct nationalities? The answer is multi-label classification, or tagging, rather than the familiar multiclass classification. Because the terms multiclass classification and multi-label classification sound so similar and are easily confused, you probably want to use the term multi-label tagging, or just tagging, instead of

    参考来源

    - 位置:《Natural Language Processing in Action, Second Edition》第08章《循环网络与长短期记忆》(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-8.html) - 结论依据:摘录中关于互斥多类与多标签目标不同的定义、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:8.2.4 Multiclass classifiers vs. multi-label taggers How can you deal with the ambiguity of surnames for which there are multiple different correct nationalities? The answer is multi-label classification, or tagging, rather than the familiar multiclass classification. Because the terms multiclass classification and multi-label classification sound so similar and are easily confused, you probably want to use the term multi-label tagging, or just tagging, instead of