887/5000
Hi everyone,
I use Ilastik, pixel classification software, for vessel segmentation. Given the labels and a set of functionalities, it uses a random forest algorithm from the VIGRA lybrary to make predictions on the pixels.
You can access the variable importance table to see which features have the most impact on the algorithm. But I don't know how to interpret it:
As you can see on the image below, there are 4 columns in the table, Class#0, Class1, Overall and Gini, with a score for each feature
What I don't understand is the Class # 0 and Class # 1 column. It looks like their sum is equal to the overall score, but I don't know if that has anything to do with the out of the bag error, which, along with, Gini importance, is often used to describe the importance of functionality in a random forest
Has anyone here encountered these class scores # 0 and # 1 for a random forest and knows what it is?
