Understanding Model

[bcarls2]

Hello, 

I am attempting to understand more about some of the models being created so I can create some simple visuals for others not close to project. My goal was to create a simple tree model to help show what the model is splitting on to get to its predictions but have not been able to find the values that help explain this or the order in which the model is going through the feature list down the length of the trees.

(for context currently trying to do this with a eXtreme Gradient Boosted Trees Classifier with Early Stopping - Forest (10x) model). 

I understand that in actuality the model is utilizing multiple different trees and finding the average prediction value between them all to get to it's final decision but just trying to see one or two of them to help provide context to people that are not super involved in Machine Learning Models. 

Thank you,

Brandan

[jenD]

Thanks for the post @bcarls2 . There is no way to do this currently, but it's a great idea. It could be done with a script to export the model and visualize the trees, but writing that script requires an understanding of DataRobot internals. I am submitting your idea to our team of data scientists who can perhaps create the script as an AI accelerator. If you haven't seen that section, there are lots of really interesting ideas, although not yours--yet!

jen