知识卡片

普通读书笔记卡 · 6653.a · 生成模型:GPT-6

字符距离与语义相似回答不同问题

理解等级:中级

Levenshtein计算插入、删除和替换等编辑成本,适于拼写比较;good与okay可以词义近而字符差很大。Jaccard比较集合重合,并非同一种逐字符编辑次数;选择距离前应先确定比较的是形式、主题还是具体意义。

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

      When we use character sequence matches to measure distance between natural language phrases, we often get it wrong. Words and phrases with similar meanings, like good and okay, often have different character sequences and large distances when we count character-by-character matches to measure distance. And sometimes, two words look almost the same but mean completely different things, like bad and bag. You can count the

    参考来源

    - 位置:《Natural Language Processing in Action, Second Edition》第01章《会读写的机器与自然语言处理》(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-1.html) - 结论依据:摘录中关于字符距离与语义相似回答不同问题的定义、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:When we use character sequence matches to measure distance between natural language phrases, we often get it wrong. Words and phrases with similar meanings, like good and okay, often have different character sequences and large distances when we count character-by-character matches to measure distance. And sometimes, two words look almost the same but mean completely different things, like bad and bag. You can count the