← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 6666.f · 生成模型:GPT-6

平滑公式应完整保存

理解等级:中级

原书图式写log(N/(1+df))+1,而scikit-learn默认平滑为log((1+N)/(1+df))+1。分子是否加1、对数底数及外加常数都会影响数值;实现时应记录实际公式和参数,不能把不同约定称为完全相同。

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

      Next you want to calculate inverse document frequency, which tells you how important the token is or how much information it adds to a documents where it is used. The most common way to estimate the information value of a token is to take the log of the ratio of the total number of documents in your corpus and the count of the documents containing

    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:摘录中关于平滑公式应完整保存的定义、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:Next you want to calculate inverse document frequency, which tells you how important the token is or how much information it adds to a documents where it is used. The most common way to estimate the information value of a token is to take the log of the ratio of the total number of documents in your corpus and the count of the documents containing