课程: Machine Learning and AI Foundations: Decision Trees with KNIME

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！

How CART handles missing data using surrogates

How CART handles missing data using surrogates - KNIME教程

课程: Machine Learning and AI Foundations: Decision Trees with KNIME

开始一个月试用 购买团队方案

How CART handles missing data using surrogates

“

- [Instructor] Let's talk about how CART handles missing input variables. CART actually has a fascinating way of addressing this. It's a technique called surrogates. Now, nine doesn't support this particular missing value strategy in its decision tree learner, but it's very interesting and I want you to be familiar with it so you'll understand how CART works on multiple platforms. So, here we go. For each split in the tree, CART identifies the input fields that are the most statistically similar to the selected split field, and we're going to have an example in a moment, but the notion is that when you're missing the input variable, let's say income, you now want to find those other variables that are correlated with income not because you're predicting income per se, but you're going to use these alternatives as a proxy. What's fascinating about this is you're not attempting to impute. In other words, you…

内容