cancel
Showing results for 
Search instead for 
Did you mean: 

Error - 'argument of length 0'

Former Member
0 Kudos

Hi

I am using Predictive Analysis 1.0.10. I can run the clustering algorithm (R-K-Means) successfully so I assume that I am set up and configured correctly, but I get an error when I run the Decision Trees algorithm (R-CNR Tree):

     "Error in 1:nrow(object$splits) : argument of length 0

I have tried various permutations of fields and settings.

Any idea what the problem might be?

Thanks

Al.

Accepted Solutions (0)

Answers (3)

Answers (3)

Former Member
0 Kudos

Thanks for your input. It works now. I didn't change anything.

Former Member
0 Kudos

Hi Al,

What is the dataset that you are using?

Thanks,

Vishwa

former_member184912
Participant
0 Kudos

Hi Al,

Do you get this error regardless of data sources used in PA?

I have 1.0.10 and R-CNR works fine here.

Try the following.

Check the logs for further information. The logs are by default are stored in %TEMP%\sapvi\logs folder.

Best regards,

Kurt Holst

Former Member
0 Kudos

Hello,

Im having the same problem, could someone please help. I have noticed that it only starts working if my 1s are in equal proportion with 0s. On my original data, i just have 8% as 1s, any ideas, please its urgent.

Looking forward to hearing from you.

former_member184912
Participant
0 Kudos

Hi Sukruth,

thanks for for reaching out.

A Way to increase your predictions is to oversample of boost your 1s.

You find this approach described here:

http://scn.sap.com/docs/DOC-50803

Good luck.

best regards,

Kurt