Alteryx Designer Discussions

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

Join us on Dec 2 for a half-day virtual analytics + data science event!
US & CA customers only

SAVE YOUR SPOT
It's the most wonderful time of the year - Santalytics 2020 is here! This year, Santa's workshop needs the help of the Alteryx Community to help get back on track, so head over to the Group Hub for all the info to get started!

Logistic Regression - applying model equation to source data

Highlighted
8 - Asteroid

I'd like to use my logistic regression equation to show which items in my source data would have been predicted to be positive, and which would have been predicted to be negative.  My target variable is: Does the employee term within 90 days?  YES/NO.  Here's the regression model formula:

 

glm(formula = TERM_90DAYS ~ DOMICILE + VET + BUS_UNIT + AGE_AT_HIRE + MAR_STATUS + SCHOOL_LEVEL + HOURLY_RT, family = binomial("logit"), data = the.data)

 

Using this formula and the coefficients, I'd like to feed in the source data to the model so I will be able to see which teammates would have been correctly predicted to term within 90 days, which would have been correctly predicted to not term within 90 days, my Type I errors, and my Type II errors.

 

So, basically, my source data has 4,058 records, and I'd like to know which of those records fall within the four quadrants shown below (from the I output on my Logistic Regression tool).  I'd like the process to be as automated as possible.  Is there an easy way to do this?

 

Capture.PNG

Highlighted
Alteryx Alumni (Retired)

Hi @oracleoftemple

 

The output you are looking for can be obtained by putting a Browse tool after the output anchor on the Logistic Regression tool. You can see a sample of this by opening the following sample workflow:

 

Help -> Sample Workflows -> Predictive Tool Samples -> Predictive Analytics -> Logistic Regression

 

Run this workflow and click the Browse tool after the I anchor.

 

image.png

 


Labels