知识卡片
层次聚类保留逐步组合关系
凝聚法从单点簇向上合并,分裂法从整体向下拆分。树状图保留嵌套关系,最终分组还需选择截断高度或簇数;完整构树到一簇不意味着业务只能保留一组。
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
- · 无监督学习模型
原书出处 · 1 处
- · 查看原书依据
The hierarchical clustering algorithm puts similar objects into groups called clusters. The outcome is a set of clusters where each cluster is distinct from each other cluster and the objects within each cluster are like each other. This method creates a hierarchy of clusters where two close clusters are in the same cluster. The algorithm ends when there is only one cluster left. The hierarchical
参考来源
- 位置:《Machine Learning Fundamentals: Concepts, Models, and Applications》第07章《无监督学习模型》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Machine Learning Fundamentals: Concepts, Models, and Applications/_epub-src/OEBPS/xhtml/20_chapter07.xhtml)
- 结论依据:摘录中关于层次聚类保留逐步组合关系的定义、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:The hierarchical clustering algorithm puts similar objects into groups called clusters. The outcome is a set of clusters where each cluster is distinct from each other cluster and the objects within each cluster are like each other. This method creates a hierarchy of clusters where two close clusters are in the same cluster. The algorithm ends when there is only one cluster left. The hierarchical