← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 9250.c · 生成模型:GPT-6

重抽后拟合模型再聚合

理解等级:中级

从训练数据有放回生成多份样本,分别拟合模型后再投票或平均,可降低部分高方差学习器的不稳定性。各模型仍共享原始数据、可能相关,聚合收益需要评价而非定义保证。

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

      The bagging method was introduced by Leo Breiman in 1996. In 1996, the method used a random sample of data in a training set. The sample is selected with replacement so that the individual data points can be selected more than once. After several samples are generated, the models are trained independently. Depending upon whether the problem is that of classification or regression, the average

    参考来源

    - 位置:《Machine Learning Fundamentals: Concepts, Models, and Applications》第06章《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/19_chapter06.xhtml) - 结论依据:摘录中关于重抽后拟合模型再聚合的定义、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 bagging method was introduced by Leo Breiman in 1996. In 1996, the method used a random sample of data in a training set. The sample is selected with replacement so that the individual data points can be selected more than once. After several samples are generated, the models are trained independently. Depending upon whether the problem is that of classification or regression, the average