Definition

$PRES S_{p} = i = 1 \sum n (y_{i} - \overset{y}{^}_{i (i)})^{2} = i = 1 \sum n (\frac{e _{i}}{1 - h _{ii}})^{2}$ where $y_{i (i)}$ is an estimated value of $y_{i}$ calculated by the $n - 1$ data excluding $y_{i}$

The predicted residual error sum of squares (PRESS) is a form of cross validation used in regression analysis