知识卡片
因果推断关注特征影响
有监督任务有时追求解释模型,目标是描述输入特征如何影响标签,而非只提高新样本标签预测准确率。
书中把这称为因果推断;建立这类解释模型需要区别于单纯拟合。
这里的“影响”是因果问题:若主动改变某个特征,标签会怎样变化。仅观察训练数据中某特征与标签相关,或让模型很好地拟合二者,并不能单独排除混杂因素、反向因果等解释,因此也不能仅凭拟合结果断言因果关系。
这张卡片看不懂?
这张卡片的反馈已满 10 条,暂时无法继续提交。
已有反馈 · 0 / 10 条
还没有人提交反馈。
在整书笔记中的位置
原书出处 · 1 处
- 第33章 机器学习算法 · 查看原书依据
Sometimes the goal in a supervised machine-learning application is not making accurate predictions of labels for new examples, but rather performing causal inference: finding an explanatory model that describes how the various features of an input data point affect its associated label. Finding a model that fits a given set of training data well can be tricky. It may involve sophisticated optimization methods that need to
参考来源
- 位置:《算法导论(第4版)》第33章《机器学习算法》(OEBPS/chapter033.xhtml,文本字符[0,5000))。
- 结论依据:原文第33章导言和33.1节的学习任务、聚类定义、目标函数与局部最优说明;公式及图已核对。
- 原始内容:Sometimes the goal in a supervised machine-learning application is not making accurate predictions of labels for new examples, but rather performing causal inference : finding an explanatory model that describes how the various features of an input data point affect its associated label. Finding a model that fits a given set of training data well can be tricky. It may involve sophisticated optimization methods that need to