cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
Try the Materials Informatics Toolkit, which is designed to easily handle SMILES data. This and other helpful add-ins are available in the JMP® Marketplace
Choose Language Hide Translation Bar

Can there be multiple classes in a dependent variable to run a decision tree model

We want to run a decision tree model to analyze survey responses and the dependent variable has multiple classes. Is it possible to run an accurate model with this or would the dependent variable have to be a binary classification

1 ACCEPTED SOLUTION

Accepted Solutions
Victor_G
Super User

Re: Can there be multiple classes in a dependent variable to run a decision tree model

Hi @PartialGopher16,

 

Welcome in the Community !

 

If I understand your question well, you would like to analyze survey responses, and each response may have different levels possible ?

Decision Tree are able to do multi-class classification. You can try on sample datasets, like Car Poll.jmp (accessible in "Help", "Sample Index", "Exploratory Modeling"), the response used in the "Partition" (decision tree) script is "Country", which has 3 possible classes : Japanese, European, American. 

The Decision Tree will use the other variables to try to predict from which country does each car owner comes from :

Victor_G_0-1712302365576.png

 

I hope I did understand your question and that this response will help you,

Victor GUILLER

"It is not unusual for a well-designed experiment to analyze itself" (Box, Hunter and Hunter)

View solution in original post

2 REPLIES 2
Victor_G
Super User

Re: Can there be multiple classes in a dependent variable to run a decision tree model

Hi @PartialGopher16,

 

Welcome in the Community !

 

If I understand your question well, you would like to analyze survey responses, and each response may have different levels possible ?

Decision Tree are able to do multi-class classification. You can try on sample datasets, like Car Poll.jmp (accessible in "Help", "Sample Index", "Exploratory Modeling"), the response used in the "Partition" (decision tree) script is "Country", which has 3 possible classes : Japanese, European, American. 

The Decision Tree will use the other variables to try to predict from which country does each car owner comes from :

Victor_G_0-1712302365576.png

 

I hope I did understand your question and that this response will help you,

Victor GUILLER

"It is not unusual for a well-designed experiment to analyze itself" (Box, Hunter and Hunter)

Re: Can there be multiple classes in a dependent variable to run a decision tree model

Thank you for your response and your help, you have answered my query