From the course: Supervised Learning Essential Training
Unlock the full course today
Join today to access over 22,600 courses taught by industry experts or purchase this course individually.
Analyzing decision tree performance - Python Tutorial
From the course: Supervised Learning Essential Training
Analyzing decision tree performance
- [Instructor] Now that we've come up with predictions, let's have a look at our tree. This is our root node of our tree because it provided us with the largest split between our samples. You can see within that node, we started with 7,500 examples, and we're able to split one group in about 5,600 examples and another into 1800. If the light wasn't over 369, then all of those samples are evaluated by their CO2 level. If we keep following down this path, we check to see if their CO2 is at a certain point. And if it's not, we check to see if the humidity is at a certain point. This then splits again down between temperature. And if we scroll all the way over here between CO2 again. All of these decisions finally lead us to leaf nodes at the very bottom. What we can see is that the Gini index is zero, which is what we want. Meaning all of the samples in each of the groups are homogenous. We can also see how many…
Practice while you learn with exercise files
Download the files the instructor uses to teach the course. Follow along and learn by watching, listening and practicing.
Contents
-
-
-
-
-
(Locked)
Identify common decision trees1m 55s
-
(Locked)
Splitting data and limiting decision tree depth3m 41s
-
(Locked)
How to build a decision tree2m 3s
-
(Locked)
Creating your first decision trees2m 49s
-
(Locked)
Analyzing decision tree performance5m 1s
-
(Locked)
Exploring how ensemble methods create strong learners1m 55s
-
(Locked)
-
-
-