知识卡片
变量定义影响可回答的问题
变量的分类、阈值与测量方式决定了分析对象。例如年龄按区间记录与连续记录提供的信息不同。应依据查询选择可解释的抽象并保存编码规则;沿用数据列名并不能保证该变量与想研究的原因或行动一致。
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
- · 构建因果图模型
原书出处 · 1 处
- · 查看原书依据
Labeling causal abstractions How we conceptualize the variables of a model matters greatly in machine learning. For example, ImageNet, a database of 14 million images, contains anachronistic and offensive labels for racial categories. Even if renamed to be less offensive, race categories themselves are fluid across time and culture. What are the “correct” labels to use in a predictive algorithm? How we define our variables
参考来源
- 位置:《Causal AI》第03章《构建因果图模型》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Causal AI/_epub-src/OEBPS/Text/chapter-3.xhtml)
- 结论依据:摘录中关于变量定义影响可回答的问题的定义、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:Labeling causal abstractions How we conceptualize the variables of a model matters greatly in machine learning. For example, ImageNet, a database of 14 million images, contains anachronistic and offensive labels for racial categories. Even if renamed to be less offensive, race categories themselves are fluid across time and culture. What are the “correct” labels to use in a predictive algorithm? How we define our variables