Separating Learning phase from Predicting phase in Alteryx

[Hagar-Usama]

Hello,

I am new to Alteryx and am planning to utilize machine learning tools in the Alteryx desktop. My goal is to design separable phases where the output of the training phase, such as a weight file, can be used independently in another workflow to predict new output. Is this possible, or is there a workaround?

[FrederikE · Accepted solution]

Hey @Hagar-Usama,

You can save your Models directly from the O-Output of the Model-Tools. 

E.g. you could save a Linear Regression Model by connecting the O-Output directly to a Output tool and save it as a .yxdb. (This would be your training step.)

This .yxdb can be called in a different workflow again. (This would be your testing step.)

[Hagar-Usama]

Perfect, that's what I was looking for. Thanks!