知识卡片

普通读书笔记卡 · 6683.a · 生成模型:GPT-6

质心平移只能校正平均域偏移

理解等级:中级

将简历与职位描述两域的平均向量差作平移,可尝试减去共同的表达偏移。它不能学习每个配对的复杂关系,若两边都同样平移,欧氏两两差反而不变;应明确平移哪一域、使用什么距离及如何验证匹配收益。

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

      One way to do this is to calculate the mean difference between your two centroids (like you did for LDA) and add some portion of this “bias” to all the resume or job description vectors. Doing so should take out the average topic vector difference between resumes and job descriptions. Topics such as beer on tap at lunch might appear in a job description but

    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:摘录中关于质心平移只能校正平均域偏移的定义、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:One way to do this is to calculate the mean difference between your two centroids (like you did for LDA) and add some portion of this “bias” to all the resume or job description vectors. Doing so should take out the average topic vector difference between resumes and job descriptions. Topics such as beer on tap at lunch might appear in a job description but