- What is an ensemble?
- Types of ensembles
- Measuring model accuracy
- Boosting, bagging, and stacking
- Visualizing bias and variance
- Important and influential ensemble algorithms
Skill Level Advanced
- [Keith] Let's face it, when you're building predictive models, one of the foremost things on your mind is model accuracy. And if you're competing to win, in a competition like those on Kaggle, you must have the most accurate of all. You know, in the past few years, just about every winning entry has been an ensemble. For a year or so, it was Random Forest. Then, XGBoost. Now, one of the hottest approaches, believe it or not, is deep stacking. There are lots of opportunities to learn coding commands to create these models, including many right here in the library.
But far too rarely do we take the time to understand what these techniques are all about. This course is a conceptual introduction to the topic. So if you use R, or Python, or a commercial package like SPSS or SAS, or any number of open-source modeling workbenches, here is your chance to really understand why ensembles are effective, and how they work under the hood. I'm Keith McCormick, and I've been building predictive models for more than 20 years. I look forward to showing you the building blocks of these techniques, and the details of the hottest ensemble algorithms.
Let's get started.