← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 6681.c · 生成模型:GPT-6

主题分配通常以词项计数为输入

理解等级:中级

词袋计数表示生成过程观察到了多少次词项,主题分配据此估计混合比例及词分布。TF–IDF的稀有度加权不是同一观测含义;预处理、停用项及n-gram会改变计数空间,文档长度也应与实际保留的词元对应。

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

      Keep in mind, you should calculate this statistic directly from your BOWs. You need to make sure you’re counting the tokenized and vectorized words in your documents and that you’ve applied any stop word filtering, or other normalizations, before you count up your unique terms. That way your count includes all the words in your BOW vector vocabulary (all the n-grams you’re counting) but only

    参考来源

    - 位置:《Natural Language Processing in Action, Second Edition》第04章《从词数到语义分析》(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-4.html) - 结论依据:摘录中关于主题分配通常以词项计数为输入的定义、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:Keep in mind, you should calculate this statistic directly from your BOWs. You need to make sure you’re counting the tokenized and vectorized words in your documents and that you’ve applied any stop word filtering, or other normalizations, before you count up your unique terms. That way your count includes all the words in your BOW vector vocabulary (all the n-grams you’re counting) but only