Mean squared prediction error
= Mean squared prediction error
{title2=$\operatorname{MSPE}$}
For a predictor $\widehat f$ and an independent response $Y^*$ at covariate $X^*$, mean squared prediction error is $\mathbb E[(Y^*-\widehat f(X^*))^2]$.
= Mean squared prediction error
{title2=$\operatorname{MSPE}$}
For a predictor $\widehat f$ and an independent response $Y^*$ at covariate $X^*$, mean squared prediction error is $\mathbb E[(Y^*-\widehat f(X^*))^2]$.