[英]Normal Equation and pinv() in Matlab/Octave
I'm studing the fundamentals of Machine Learning and, when I was reading about the use of Normal Equation in Matlab, I had a question.我正在研究机器学习的基础知识,当我阅读有关在 Matlab 中使用正态方程的信息时,我有一个问题。
If the Normal Equation is:如果正规方程是:
And和
Is the Pseudo Inverse , so:是伪逆,所以:
And pinv(X)
compute the pseudo inverse of X
.并且
pinv(X)
计算X
的伪逆。
Why, in Matlab/Octave, some people use:为什么,在 Matlab/Octave 中,有些人使用:
theta = pinv(X'*X)*X'*y ...(4)
instead of代替
theta = pinv(X)*y ...(5)
? ?
Is (4) not redundant? (4) 不是多余的吗? doesn't
pinv(X)
compute the pseudo inverse formula for us? pinv(X)
不会为我们计算伪逆公式吗?
Thank you!谢谢!
Has anybody found an answer to this?有人找到答案了吗? Even here under "Implementation" they use (4): http://mlwiki.org/index.php/Normal_Equation#Implementation
即使在“实施”下,他们也使用(4): http://mlwiki.org/index.php/Normal_Equation#Implementation
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.