In machine learning we may not always have a single high-performing model to make predictions with, and may instead have a set of lower-performing models. We can still run each of these lower-performing models and combine them to produce a better result than if they were run individually.
Taking an ensemble approach to decision making is equivalent to asking a range of people for their opinions: each person has their own model of the world, and the advice will therefore be different. After receiving advice, we may even find better questions to ask, to yield even more useful information. At a minimum, we are able to gather a richer set of perspectives than if we’d just asked a single person or relied on our own experience.
Return home.