机器学习笔记之正规矩阵的不可逆性
2017-07-10 08:39
363 查看
Normal Equation Non-invertibility
When implementing the normal equation inoctave we want to use the 'pinv' function rather than 'inv.' The 'pinv'function will give you a value of θ evenif XTX is
notinvertible.
If XTX is non-invertible, thecommon
causes might be having :
· Redundant features, where two features are very closelyrelated (i.e. they are linearly dependent)
· Too many features (e.g. m ≤ n). In this case, delete somefeatures or use "regularization" (to be explained in a later lesson).
Solutions to theabove problems include deleting a feature that is linearly dependent withanother or deleting one or more features when there are too many features.
When implementing the normal equation inoctave we want to use the 'pinv' function rather than 'inv.' The 'pinv'function will give you a value of θ evenif XTX is
notinvertible.
If XTX is non-invertible, thecommon
causes might be having :
· Redundant features, where two features are very closelyrelated (i.e. they are linearly dependent)
· Too many features (e.g. m ≤ n). In this case, delete somefeatures or use "regularization" (to be explained in a later lesson).
Solutions to theabove problems include deleting a feature that is linearly dependent withanother or deleting one or more features when there are too many features.
相关文章推荐
- 机器学习笔记(三)矩阵和线性代数
- 程序员的机器学习入门笔记(三):数据挖掘中矩阵的那些事
- Andrew Ng 机器学习笔记 06 :正规方程
- 机器学习笔记_数学基础_5-矩阵理论
- 【ML学习笔记】6:机器学习中的数学基础6(对角矩阵,对称矩阵,正交矩阵,特征分解)
- 吴恩达 机器学习笔记三(lecture 3)(矩阵和向量)
- 七月算法机器学习笔记2--机器学习中的数学之矩阵分析与应用
- 机器学习笔记之正规方程
- 机器学习笔记(2)---监督学习之正规方程
- Andrew Ng 机器学习笔记 03 :线性代数之矩阵
- 斯坦福大学机器学习笔记——特征和多项式回归以及正规方程
- 七月算法机器学习笔记3 线性代数与矩阵
- 【学习笔记】斯坦福大学公开课(机器学习) 之二:正规方程
- 机器学习笔记一二 - 线性规划 梯度下降 正规方程
- 【ML学习笔记】4:机器学习中的数学基础4(线性变换的矩阵描述)
- [机器学习] UFLDL笔记 - Convolutional Neural Network - 矩阵运算
- 机器学习笔记(一):梯度下降算法,随机梯度下降,正规方程
- 【机器学习笔记】Hessian矩阵
- 机器学习笔记_数学基础_5-矩阵理论_续1_QR分解
- 机器学习笔记(二)矩阵和线性代数 例:用Python实现SVD分解进行图片压缩