From the course: Applied Machine Learning: Ensemble Learning (2022)

Unlock the full course today

Join today to access over 24,700 courses taught by industry experts.

What is stacking?

What is stacking? - Python Tutorial

From the course: Applied Machine Learning: Ensemble Learning (2022)

Start my 1-month free trial Buy for my team

What is stacking?

“

- [Instructor] Now let's talk about the last of the three ensemble techniques that we'll cover in this course. And that's stacking. Stacking is an ensemble method that creates one strong metamodel that's trained on the predictions of several independent base models. On the surface, that may sound awfully similar to boosting and bagging. And the reality is that these are similar since they all fall under the umbrella of ensemble learners. So let's dig into the ways in which stacking is different than both boosting and bagging. The first is that the base models and metamodel are all trained on the same dataset, the full training set. Remember that in boosting and bagging, we use sampling techniques to generate the data for our base models to train on. In part, that's to make sure each model is different from each other and keying in on different trends in the data. We don't need to do that with stacking because the…

Contents