← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 6481.a.4.a · 生成模型:GPT-6

二值与频次保存不同信息

理解等级:中级

把到访记录编码为0/1,保留是否出现却丢掉重复频次;用计数则保留强度,也更易受日常必经地点影响。二者对应不同观察假设,不能把频次高直接当作满意度高,也不能把未观测零视为确定不喜欢。

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

      it may ignore the frequency of check-in. Few users visit their preferred location. Therefore, many entries in the matrix are left with ‘0’s. It creates data sparsity, and to overcome such a problem, both the user-based and item-based POI approaches are applied. Various studies [29] found that user-based POI recommendation is better than item-based one, which leads to inaccurate item similarity as compared with user

    参考来源

    - 位置:《Artificial Intelligence and Data Science in Recommendation System: Current Trends, Technologies, and Applications》第11章《张量分解的兴趣点推荐》(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-C11.html) - 结论依据:摘录中关于二值与频次保存不同信息的定义、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:it may ignore the frequency of check-in. Few users visit their preferred location. Therefore, many entries in the matrix are left with ‘0’s. It creates data sparsity, and to overcome such a problem, both the user-based and item-based POI approaches are applied. Various studies [29] found that user-based POI recommendation is better than item-based one, which leads to inaccurate item similarity as compared with user