When building a decision tree model, one used the GridSearchCV$()$ method to search for optimal parameters, i$.$e$.,$ 'max$\_$depth' and 'min$\_$samples$\_$split'. He set the first set of parameters as the follows:

param$\_$grid $=\{$

'max$\_$depth': $[5,10,15,20,25],$

'min$\_$samples$\_$split': $[10,20,30,40,50,60,70,80,90,100]$

$\}$

The first attempt of GridSearchCV$()$ produces the following best parameters: Initial parameters: $\{\text{'}$max$\_$depth': $5,$ 'min$\_$samples$\_$split': $40\}$

How should the person set the second param$\_$grid? Why?