知识卡片

普通读书笔记卡 · 8858.c · 生成模型:GPT-6

特征矩阵保留样本与维度职责

理解等级:中级

特征矩阵Φ每行对应一个样本,每列对应一个基函数。多项式K阶会提供K+1列,包含常数列;正规方程形式为ΦᵀΦθ=Φᵀy。逆矩阵解仍需列满秩,样本数少于列数或基函数相关时不能未经检查直接求逆。

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 matrix Φ is an array of the features (columns) for each datapoint (rows). It plays the same role as data matrix X did in chapter 9; in fact we often call matrices of the feature representations of each datapoint also as a data matrix. As an exercise, you can derive the closed form of the optimizer, given above, using the same steps as we

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第11章《泛化问题》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)/_epub-src/OPS/xhtml/c011.xhtml) - 结论依据:摘录中关于特征矩阵保留样本与维度职责的定义、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 matrix Φ is an array of the features (columns) for each datapoint (rows). It plays the same role as data matrix X did in chapter 9; in fact we often call matrices of the feature representations of each datapoint also as a data matrix. As an exercise, you can derive the closed form of the optimizer, given above, using the same steps as we