Consider a scenario where you have to predict the likelihood of a customer churning. You are given three different models: a decision tree, a logistic regression, and a neural network. Which ensemble method would you use to combine these models, and why?