课程: Advanced Predictive Modeling: Mastering Ensembles and Metamodeling

今天就学习课程吧!

今天就开通帐号,24,600 门业界名师课程任您挑!

What is Bootstrap aggregating?

What is Bootstrap aggregating?

- [Instructor] The concept of bootstrap aggregating which you've probably heard called bagging goes all the way back to the '90s and it was first proposed by Leo Breiman. He's famous for a few things. But three things in particular that are of interest to us. Cart, bagging and random force. In one of the early papers on the technique he explains that it involves multiple versions of a model and then you just average 'em if it's a scale or you do a plurality vote when you're predicting a category but here's the key point. The multiple versions are formed by making bootstrap replicates of the learning set. But what's a bootstrap replicate? Let's take a look at an example. I've got an Excel spreadsheet with the auto data and I've generated 100 bootstrap samples. So, let's take a closer look. We've got our Chevy Malibu appears once, we've got our Dodge Challenger shows up a few times actually so what's going on? When you do bootstrap sampling, you're choosing a case at random but then…

内容