Algorithm 3 Experiments For Hyperparameter Selection

$results\leftarrow \left[\right]$ for $final\_prec$ in 0.6, …, 0.95 do     for f in 0.1, …, 0.5 do         $newset\_prec\leftarrow$ select the $des\_feat\_num$ of number of features from the original set with $final\_prec$ and f         results.append( $newset\_prec$, $newset\_psovft$) find the best performance in $results$ and take the corresponding hyperparameters