知识卡片

普通读书笔记卡 · 8858.b · 生成模型:GPT-6

基函数回归对参数仍是线性

理解等级:中级

多项式回归可写fθ(x)=θᵀφ(x),其中φ(x)=[1,x,…,xᴷ]ᵀ。它对原始x可以非线性,对待学系数θ却是线性,因此可在特征矩阵上使用线性最小二乘方法。此技巧也适用于其他固定基函数;若基函数参数同时学习,问题结构还会改变。

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

      which is a linear function in the parameters θ. Therefore, if we featurize x, representing each datapoint x with a feature vector ϕ(x), then we have arrived at a linear regression problem in this feature space. So, the learning problem, and closed form optimizer, for L2 polynomial regression looks almost identical to that of L2 linear regression:

    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:摘录中关于基函数回归对参数仍是线性的定义、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:which is a linear function in the parameters θ. Therefore, if we featurize x, representing each datapoint x with a feature vector ϕ(x), then we have arrived at a linear regression problem in this feature space. So, the learning problem, and closed form optimizer, for L2 polynomial regression looks almost identical to that of L2 linear regression: