Pulling data from Linear Regression report output

I'm using the Linear Regression tool and would like to continue to use some of it's output. Specifically, fields from the coefficients table such as intercept. Is there a way to do this?

Predictive Analysis

Accepted answers

rahuls

Hi

As per your query, what you can do is:
1. Open the Linear Regression Tool and Goto R tool.
2. From the first output (1) of R tool, you can parse the data (using Text to Columns Tool) and get the required output you would like to continue with.
3. For furthur information, you can check the following snapshot of customized Linear Regression. The highlighted area is the area of your interest.

.

Hope this will help you.

All comments

rahuls

Hope this will help you.

FDELABAL

Hi,

I'm confused as to how this works specifically. What are you parsing by when you add the text/column function?? The report that comes out from that output is not broken down in a way where you can easily parse out what variable is tied to what coefficient, etc.

This seems like a pretty basic functionality that's missing in Alteryx. You would think it's obvious that we want an easy way to use the regression output for further calculations (e.g., predicting what model says would be variable given x, y, z, etc.) Am I missing something?

chris_love

Hold on guys. There's a simple macro you can use without parsing the reports. It uses R code to flesh out the R model from the model output of any of the predictive tools. Its written to address exactly the issue the OP raised.

I'm not at my PC right now but I'll post it first thing (its v late here in UK).

chris_love

I have this macro in my old email address which has gone, but the solution is akin to this one:

http://community.alteryx.com/t5/Advanced-Analytics/Spline-Tool-Output/td-p/325

Perhaps @DrDan can post the relevant R code we can turn into a macro as this link doesn't work for Linear Regression.

Chris

mseelan

Hi @DrDan , this was truly very useful, how do we retrieve the fitted values from the linear regression model? help with this would be much appreciated

Thanks

Mark

Quick Links

This months top contributors

atcodedog05 19598

Qiu 15875

binu_acs 15708

MarqueeCrew 13708

apathetichell 13703