In $2$D scatter plots, you are given points from $2$ classes, shown as $+$s and $-$s$.$ For each of these sets of points,

$1.(2\mathrm{.}5\%)$ Draw the decision tree of depth at most $2$ that can completely separate the given data in each of the figures, by filling in the predicates $($only involving the thresholding of a single variable$)$ in the boxes for the decision trees below. Recall how we did it in the class. If the data is already separated when you hit a box in your tree, then write the class $($i$.$e$.,$Plus$,$ or $$Minus$),$ and leave the sub$-$tree unfilled from that box empty. In $2$D scatter plots, you are given points from $2$ classes, shown as $+\text{'}$s and $-\text{'}$s$.$ For each of these sets of points,

$1.(2\mathrm{.}5\backslash \%)$ Draw the decision tree of depth at most $2$ that can completely separate the given data in each of the figures, by filling in the predicates $($only involving the thresholding of a single variable$)$ in the boxes for the decision trees below. Recall how we did it in the class. If the data is already separated when you hit a box in your tree, then write the class $($i$.$e$.,$ "Plus", or "Minus"$),$ and leave the sub$-$tree unfilled from that box empty. $2.(2\mathrm{.}5\backslash \%)$ Draw splits on the scatter plots and write the class labels for each of the resulting bins somewhere inside the resulting bins $($such as "Plus" or "Minus". If you think the data in any of these examples cannot be completely separated by a depth $2$ tree, then do not draw anything on it and simply indicate "Inseparable data".