知识卡片
前排优化与全列表优化有别
只展示前几个结果的场景更关心列表前排,而为任意类别子集排序又有不同覆盖要求。优化目标和评价截断长度应匹配真实展示位置;前排算法不等于只能读取原候选列表头部,仍需适当候选生成。
这张卡片看不懂?
这张卡片已通过人工审查,不再接受失败反馈。
已有反馈 · 0 条
还没有人提交反馈。
在整书笔记中的位置
- · 电影推荐
原书出处 · 1 处
- · 查看原书依据
its prime focus is on the head of the ranking. It is impossible with PVR as in a catalog; ranking arbitrary subsets is done using it. The top N ranker is an optimized algorithm tested according to several metrics. It searches only the head of the catalog produced by the algorithm instead of working on the entire one.
参考来源
- 位置:《Artificial Intelligence and Data Science in Recommendation System: Current Trends, Technologies, and Applications》第06章《电影推荐》(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-C6.html)
- 结论依据:摘录中关于前排优化与全列表优化有别的定义、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:its prime focus is on the head of the ranking. It is impossible with PVR as in a catalog; ranking arbitrary subsets is done using it. The top N ranker is an optimized algorithm tested according to several metrics. It searches only the head of the catalog produced by the algorithm instead of working on the entire one.