知识卡片

普通读书笔记卡 · 6487.g · 生成模型:GPT-6

跨域评分归一不消除语义差异

理解等级:高级

不同域可能记录点赞、切换行为或五星评分。数值归一有助于对齐范围,却不能把这些信号自动变成相同强度的偏好。迁移前应说明反馈类型与含义;本章为形式化方便假设同一五星范围,应用时须检查是否成立。

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

      TV domain, the user feedback could be implicit (e.g., a user switches for another TV show within a few minutes/seconds watching it) or explicit (e.g., “Like” or “Dislike”). Whereas 5-star or even 10-star scales could represent music and book ratings. In that situation, normalization strategies could be used [73]. We should consider user ratings on a 5-star scale for convenience by considering the presented formalization.

    参考来源

    - 位置:《Artificial Intelligence and Data Science in Recommendation System: Current Trends, Technologies, and Applications》第04章《多领域情境推荐的张量分解》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Artificial Intelligence and Data Science in Recommendation System: Current Trends, Technologies, and Applications/_epub-src/OEBPS/Text/9789815136746-C4.html) - 结论依据:摘录中关于跨域评分归一不消除语义差异的定义、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:TV domain, the user feedback could be implicit (e.g., a user switches for another TV show within a few minutes/seconds watching it) or explicit (e.g., “Like” or “Dislike”). Whereas 5-star or even 10-star scales could represent music and book ratings. In that situation, normalization strategies could be used [73]. We should consider user ratings on a 5-star scale for convenience by considering the presented formalization.