community
cancel
Showing results for 
Search instead for 
Did you mean: 

Alteryx designer Discussions

Find answers, ask questions, and share expertise about Alteryx Designer.

Boolean Type - Errors in Predictive Models

Highlighted

I'm experiencing issues with Boolean Data Type variables Predictive models. I'm using the titanic data set from Kaggle to play around with Alteryx. 

 

When I set "Survived" as a string, I'm able to run the data through Decision Tree, Logistic Regression, Naïve Bayes Classifier, Boosted Model and Forest Model without any issues.

 

When I change  "Survived" to Boolean value, Logistic Regression and the Naïve Bayes Classifier both refuse to acknowledge the variable as a target option.  The Decision Tree and Forest Model will both run, but with the warning that "The target variable is numeric, however, it has 4 or fewer unique values." The Boosted model gives the error "The minimum of the loss function was likely not obtained."

 

Can someone explain to me why the Predictive models are able to proceed with the binary option in a string, but can't as a Boolean?

 

Thanks,

 

M. Kilfoil

 

 

Community Content Engineer
Community Content Engineer

Hi @M_Kilfoil - yes, the R tools traditionally do not "allow" Boolean types - you can see this when you open up the tools and check the X_vars and Y_vars configurations:

 

X_vars config.JPG

 

I believe this is due to how R handles Boolean values, but perhaps some of the more knowledgeable Community users will have more specifics...?

 

http://stackoverflow.com/questions/5681166/what-evaluates-to-true-false-in-r

https://www.r-bloggers.com/data-types-part-4-logical-class/

Labels