error in boosted and forest models?!

Dear all,

I have been working on this project for a while, and i kept stumbling upon a problem that tired me.

when i run my workflow, i receive 6 errors:

boosted:

Error: Boosted Model (28): Boosted Model: Error in checkForRemoteErrors(val) :
Error: Boosted Model (28): Boosted Model: Execution halted
Error: Boosted Model (28): Boosted Model: The R.exe exit code (1) indicated an error.

forest:

Warning: Forest Model (35): 350 records were omitted due to missing values
Error: Forest Model (35): Forest Model: Error in randomForest.default(m, y, ...) :
Error: Forest Model (35): Forest Model: Execution halted
Error: Forest Model (35): Forest Model: The R.exe exit code (1) indicated an error.

Keeping in mind that i did clean up the data, and imputed some.

can someone please help me on this matter?

I will attach my workflow below.

thank you

-Dalal

Project 4 fixed.yxmd

Help

Accepted answers

danilang

Hi @dalilo99

Your Imputation tool is configured to Output Imputed Values as a separate field. This gives you 3 age fields that you're passing to the predictive tools

Age-years is the original field and the imputed age column( with the hyphen) is the field that shows value replaced with imputed values and the second one is the output field. The first field is completely blank and that's causing both the Boosted and Forest model to choke. Remove this one from the predictor fields in the models. Of the two remaining age fields, you should only use one, because they're redundant.

Also...In your imputation tool, you're replacing nulls with the median value but there are no nulls in your data set. Logically you should be replacing the 0 ages with the median, since 0 is not a valid value for age.

Dan

All comments

danilang

Hi @dalilo99

Can you attach the input file, "Credit-data-training.xlsx"?

Dan

dalilo99

Sure

credit-data-training.xlsx

Quick Links

This months top contributors

LanisC 335

hhaigh1 280

FláviaB 262

JORGE4900 235

caltang 177