← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 6665.a · 生成模型:GPT-6

二元出现与词项次数不同

理解等级:中级

逐位置独热向量按列取逻辑或得到是否出现,按列求和得到出现次数。前者无法区别一次和多次出现,后者仍丢失词序;需根据任务选择,不能将词项集合、计数向量和原始序列混为一谈。

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

      Let’s dig deeper into the challenge of representing a piece of text as a numerical vector that a machine can work with. In the previous chapter, you created your first vector space model of a text. You used one-hot encoding of each word and then combined all those vectors with a binary OR (or clipped sum) to create a vector representation of a text. And

    参考来源

    - 位置:《Natural Language Processing in Action, Second Edition》第03章《词频与逆文档频率向量》(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-3.html) - 结论依据:摘录中关于二元出现与词项次数不同的定义、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:Let’s dig deeper into the challenge of representing a piece of text as a numerical vector that a machine can work with. In the previous chapter, you created your first vector space model of a text. You used one-hot encoding of each word and then combined all those vectors with a binary OR (or clipped sum) to create a vector representation of a text. And