While the developers have attempted to make their product easy to use and have built in a large number of convenience features, as with any statistical method, problems can arise. Here we have attempted to provide answers to the most frequently-asked questions.
CART FAQs
- Q1. What is CART?
- Q2. What makes Salford Systems' CART the only "true" CART?
- Q3. What is a decision tree?
- Q4. What makes CART so easy to interpret?
- Q5. How are decision trees grown?
- Q6. Why is CART unique among decision-tree tools?
- Q7. What tree-growing, or "splitting", criteria can CART provide?
- Q8. What are "adjustable misclassification penalties"?
- Q9. What are "intelligent surrogates for missing values"?
- Q10. What are CART's "automatic self-validation procedures"?
- Q11. What is a "multiple-tree, committee-of-expert method", or "bootstrap aggregation"?
- Q12. When can CART be used to advantage as a stand alone package?
- Q13. How can CART complement other data mining packages and/or suites?
- Q14. How quickly can CART generate results?
- Q15. COMBINE (Bagging, ARCing) Command
- Q16. No Tree Built
- Q17. Insufficient Memory
- Q18. High Costs, or All Relative Costs Exceed One
- Q19. Cross-Validation Breakdown
- Q20. Costs are Too High
- Q21. Cannot Apply Tree to New Data
- Q22. Too Many Levels in a Categorical Predictor
- Q23. The Tree is Too Large or Complex
- Q24. Limitations on the Learn Sample Size When Using Cross Validation
- Q25. Interactive Splitting
- Q26. Internal Error Messages
- Q27. CART and Large Datasets
- Q28. Why does the tree change when non-splitting variables are dropped?
- Q29. Improvement Penalties
- Q30. Variable importance
- Q31. Nominal (ordered) Variables and Rank
- Q32. Cross Validation
- Q33. The Systat Dataset Format
MARS FAQs
- Q1. What is MARS?
- Q2. How does MARS help analysts with regression modeling?
- Q3. How does MARS differ from conventional regression?
- Q4. How does MARS construct its models?
- Q5. What control over modeling does MARS provide the user?
- Q6. How does MARS handle missing values?
- Q7. How does MARS ensure that a model will perform as claimed on future data?
- Q8. How can MARS models be implemented for predictive purposes?
- Q9. How does MARS compare with neural nets?
- Q10. Why is MARS better than a decision tree for regression?
TreeNet® FAQs
- Q1. How does TreeNet® fit into the scheme of data mining tools?
- Q2. How does TreeNet® work and what does a TreeNet® model look like?
- Q3. What are the advantages of TreeNet® ?
- Q4. What does TreeNet® output look like?
- Q5. How are prediction and scoring handled?
- Q6. What are the advantages of TreeNet® over a neural net?
- Q7. Can a neural net do anything a TreeNet® cannot?
- Q8. What is the technology underlying TreeNet® and how does it differ from boosting?
- Q9. What is the TreeNet® track record?
- Q10. How does TreeNet® fit into the Salford Systems data mining solution?
- Q11. What are the hardware and resource requirements of TreeNet® ?
- Q12. Where can I learn more about TreeNet® ?
RandomForests FAQs