Neural Sync Active

Variable Selection

Registry Synced

76 words

1 min read

Reading compass

Now · 7.1 Selection Criteria

7.1 Selection Criteria

Criterion	Formula	Lower is
$C_p$ (Mallows)	$C_p = \frac{SS_{res}(p)}{ \hat{\sigma}^2} - n + 2p$	Better (close to $p$ )
AIC	$-2\ln(L) + 2p$	Better
BIC	$-2\ln(L) + p\ln(n)$	Better (penalizes complexity more)
Adjusted $R^2$	$1 - \frac{SS_{res}/(n-p)}{SS_{tot}/(n-1)}$	Higher is better

Forward selection: Start empty, add variables one by one
Backward elimination: Start full, remove least significant
Stepwise: Both forward and backward Join Discord PreviousTransformationsNextGeneralized Linear Models