I have a data with 12 predictor variables and one class variable. In the our method, I had to prune the decision tree with the best cp criteria. However, my output showed that there is just a "single terminal node" as the final result. What is the rule for the "single terminal node".? How can I express it as rule?
This is exactly the question I have to answer.
Take a look at this book it is by the best people in the field:
http://www-bcf.usc.edu/~gareth/ISL/
Best wishes, David Booth
PS on the last go around I missed the multivariate reference. I apologize. DB
Your algorithm decided that from the 12 predictor variables only one is needed. From the picture, it's unclear what is your class label. The rule from your tree is: if ontime then class A, else class B. This is with >80% confidence based on the training data.
- Badges
- Science topic
- Similar topics
- Organizational Psychology
- Training