Mallow's Cp

Definition

$C_p=\frac{SS{E}_p}{s^2}-(n-2p)=\overline{\text{err}}+\frac{p}{n}{\hat{\sigma }}^2$ where $p-1$ is the number of explanatory variables, $s^2$ is a sample variance under the full model, and $\overline{\text{err}}=\frac{RSS}{n}$

Mallow’s $C_p$ is used to assess the fit of a regression model. A small value of $C_p$ means that the model is relatively precise.