Stepwise Error with Logistic Regression

[kwkeb]

Hello,

I'm new to Alteryx, so forgive me if this has been addressed before. I am following the Udacity Binary Classification course on the Alteryx website, where we're building a logistic regression model and the using the Sepwise tool. However, I get the following error: Error: Stepwise (11): Tool #5: Error in parse(text = paste(orig.name, "<- orig.model")) :

I'm completely clueless as to what this means. There are no missing data in the dataset. I've taken pains to assign variables their right types. I am aware that the Stepwise tool requires two inputs: one from the model output (in this case logistic) and one from the data that was supplied to the model. These were done. Also I have attached the yxmd and the dataset as well. Any help will be appreciated. Thank you.

[kwkeb]

Sorry for any confusion. The dataset is actually attached as well. Thanks

[CristonS · Accepted solution]

Hi @kwkeb - great job on the data investigation!  That's always the first step.

In your case, please rename your model with no spaces (I used Logistic_Regression).  Per the Help page, model names must start with a letter and may contain letters, numbers, and the special characters period (.) and underscore (_). No other special characters are allowed (including spaces), and R is case sensitive.

Thanks,

Criston

[kwkeb]

Hi Criston - thanks for the response. It solved it! I can't believe something like an underscore would trip me. Thanks again!

[navneett]

This solution did not work for me. I am still getting this error. What can be the reason?

Name of the model - Linear_Regression_New

Stepwise (9) Tool #5: Error in parse(text = paste(orig.name, "<- orig.model")) :

[CristonS]

Hi @navneett, these are very difficult to diagnose without seeing your data and workflow configuration.  Can you send a yxzp?  Options> Export Workflow

Thanks.

[Gioprava]

I had the same problem with Alteryx 11.0: had to rename everything without any underscore and then it worked.

Giovanni

[CristonS]

Hi @Gioprava - which tools were you using?  Linear Regression 1.1 and Stepwise?  And removing the underscores resolved the error? 

[Gioprava]

Hi, yes, I was following the instructions from the video in the Alteryx nanodegree course from Udacity, and named everything the same way, but it wasn't working.

The only thing I changed was removing the underscores and everything went fine.

Now I'm a little puzzled because as a test I renamed everything with underscores and it's working....  

Giovanni

[yclian]

It's not weight -- that's the expected behavior. Take note that, you don't just fix the StepWise function name, you also need to fix the name of upstream function that is linked to it, e.g. your Regression function.