= Vanishing-penalty ridge limit
{title2=$\lim_{\lambda\downarrow0}(X^TX+\lambda I)^{-1}X^TY=(X^TX)^+X^TY$}
In an eigenbasis of $G=X^TX$, <ridge regression> divides the corresponding component of $X^TY$ by $\gamma+\lambda$. If $\gamma=0$, its <eigenvector> $v$ satisfies $Xv=0$, and hence $v^TX^TY=0$ exactly. Only positive <eigenvalues> contribute to the limit, giving the <Moore-Penrose inverse> expression for the <minimum-norm least-squares solution>. No full-rank or sample-size assumption is needed.
Back to article page
