知识卡片

普通读书笔记卡 · 9382 · 生成模型:GPT-6

学习预测关系不等于识别因果

理解等级:中级

模型可从输入输出样例学到可用于预测的统计规律,无需先手写全部公式。但数据选择与反馈仍影响结果,相关规律也不自动代表因果;业务问题和使用方式需要先定义,不能让拟合分数替代目标。

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

      With machine learning, you don’t have to understand how to turn the inputs into outputs—you may not even have any idea what the underlying physical relationship is between them. For home prices, you may not understand how the 100 or so features listed on a real estate advertisement are related to the price. Is it the swimming pool, the tile roof, or perhaps the nearby

    参考来源

    - 位置:《Natural Language Processing in Action, Second Edition》附录D《附录D:机器学习工具与技术》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Natural Language Processing in Action, Second Edition/_epub-src/OEBPS/Text/appendix-d.html) - 结论依据:摘录中关于学习预测关系不等于识别因果的定义、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:With machine learning, you don’t have to understand how to turn the inputs into outputs—you may not even have any idea what the underlying physical relationship is between them. For home prices, you may not understand how the 100 or so features listed on a real estate advertisement are related to the price. Is it the swimming pool, the tile roof, or perhaps the nearby