Decision tree uses trees to predict where each nodes are created by calculating entropy and information gain of all possible independent variables. The one with highest information gain is choosen as that node.