Random Forest Model:

$$ Build a Random Forest model.

$$ Tune hyperparameters such as max$\_$depth, min$\_$samples$\_$split,

n$\_$estimators, and min$\_$samples$\_$leaf using Grid Search.

$3.$ K$-$Nearest Neighbors Model:

$$ Build a K$-$Nearest Neighbors $($KNN$)$ model.

$$ Tune hyperparameters such as n$\_$neighbors, weights, and p $($dis$-$

tance metric$)$ using Grid Search.

Which metric, accuracy or F$1-$score, would you use to run the Grid$-$

SearchCV?

Select the metric $($accuracy or F$1-$score$)$ that you believe best suits

your use case.

Below are the grids for GridSearchCV:

param$\_$grid$\_$dt $=\{$

'max$\_$depth': $[3,5,10,$ None$],$

'min$\_$samples$\_$split': $[2,10,20],$

'min$\_$samples$\_$leaf': $[1,5,10]$

$\}$

param$\_$grid$\_$rf $=\{$

$\text{'}$n$\_$estimators': $[50,100,200],$

'max$\_$depth': $[$None$,10,20],$

'min$\_$samples$\_$split': $[2,10],$

'min$\_$samples$\_$leaf': $[1,5]$

$\}$

param$\_$grid$\_$knn $=\{$

$\text{'}$n$\_$neighbors': $[3,5,7,9],$

'weights': $[\text{'}$uniform$\text{'},$ 'distance'$],$

$\text{'}$p$\text{'}$: $[1,2]$

$\}$

Decision Tree $($DT$)$ Grid:

$$ max$\_$depth: Controls the maximum depth of the tree. Options: $3,$

$5,10,$ or no limit $($None$).$

$2$