← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 9387 · 生成模型:GPT-6

回归误差和拟合优度回答不同问题

理解等级:中级

RMSE用目标同单位概括误差大小,R²比较残差平方和与以真实均值为基线的平方和。R²可为负,不是一般的皮尔逊相关系数平方;跨任务分数也受目标方差与数据难度影响,不能直接排名能力。

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

      D.4.2 Measuring regressor performance The two most common performance scores used for machine learning regression problems are root mean squared error (RMSE) and the coefficient of determination (R2). RMSE is the most useful option for most problems because it tells you how far away from the truth your predictions are likely to be. Listing D.7 RMSE >>> y_true = np.array([0, 0.5, 0.8, 0.1, 1, 0.9,

    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:摘录中关于回归误差和拟合优度回答不同问题的定义、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:D.4.2 Measuring regressor performance The two most common performance scores used for machine learning regression problems are root mean squared error (RMSE) and the coefficient of determination (R2). RMSE is the most useful option for most problems because it tells you how far away from the truth your predictions are likely to be. Listing D.7 RMSE >>> y_true = np.array([0, 0.5, 0.8, 0.1, 1, 0.9,