It turns out that enemies move around the world, so a simple K$-$Nearest$-$Neighbors on

their location won't do a very good job of protecting your creature. Instead, you decide

to use a Decision Tree, based on some characteristics of the creatures in this world, to

classify creatures as Enemies or not $($Enemy$=$Y$/$Enemy$=$N$).$ The other $5$ characteristics

$($Character$,$ Talk, Boss, Annoying and Killable with Jump$)$ are listed also in the table

below.

Consider this as a training set, draw a decision tree for this training set $($you don$$t need

to compute information gain for Character$)$