Linear Regression Tool: Data Extraction

Question

Hello

I am using the Linear Regression tool and am wondering if there is a way/solution to extract the predicted data points (as seen in the model performance chart below) in a consolidated output rather than going point by point and extracting manually?

Ideally the solution would be able to align the actual values against the predicted values in side by side columns.

AngelosPachis · Accepted Answer

Hi @JK15 ,

The values you can see in the Model Performance tab of the Interactive output anchor are essentially the values you would get if you apply the predictive model you have created with the Linear Regression tool to the dataset you've used to train that model.

To get this output, all you have to do is use a Score tool, connect the O output anchor of the Liner Regression tool to the M input anchor of the Score tool and create a connection between the dataset you used to train your model to the D input anchor of the score tool.

I used the sample Alteryx workflows to mock up an example for you. What I described above is essentially this:

In the Interface output, I have displayed two data points as displayed below

which you can identify in the score tool output

However, be worry of model bias when using the training dataset to make predictions; obviously the model will appear to be performing better when the values it's called to predict are those used to train it.

Hope that helps,

Angelos