Cross validation under decision tree

Question

I tried to use the cross-validation setting under decision tree, however, when I took a look at the confusion matrix, it is still for the original setting 70/30 split. I attached the workflow.

I saw that there is a cross validation tool but I could not find it under predictive analytics tab.

Thank you.

diabetes CV.yxzp

lixuanzhang · Answer

I now have trouble figuring out how the tool works. I would like to do 10 fold so 90% of the data will be used as training data and 10% of data will be used as testing dataset then this should be repeated 10 times. So every record in the dataset will be used in the testing dataset. However,  the cross-validation needs input from the decision tree first, and in my workflow, the decision tree used 100% of the data. I tried to use create sample tool, but I cannot figure out how to connect which tools to which tools.

I attached the file. The AUC values are below 0.5 which is not correct.

diabetes CV.yxzp

lixuanzhang · Answer

Thanks! I will try it out.

Garabujo7 · Answer

Hello, you can download the crossvalidation tool from the Alyteryx community.

https://community.alteryx.com/t5/Public-Community-Gallery/Cross-Validation-Tool/ta-p/878754

It's free, just download it and install it.

If you'd like to add the macro to your predictive palette, here are the instructions as well.

https://community.alteryx.com/t5/Alteryx-Designer-Discussions/Add-Macro-to-toolbar/m-p/184976#M47813

Gabriel