Tool Icon

Machine Learning Predict

Last modified: July 18, 2023

Use the Machine Learning Predict tool to use models you've built in Alteryx Machine Learning to make predictions about new data.

The Machine Learning Predict tool requires the AMP engine to run. Note that AMP is the default engine for all new workflows starting with Designer 22.1. Learn more about the AMP engine.

Install the Tool

To install the Machine Learning Predict tool in Designer...

  1. Go to the downloads page.
  2. Go to Alteryx Intelligence Suite > Alteryx Intelligence Suite (your Designer version).
    1. Note that you must have Designer 21.3 or higher installed.
  3. Download the Alteryx Machine Learning Designer Integration Tools YXI file.
  4. Open Designer with the "Run as Administrator" option enabled.
  5. Open the YXI file to start the installation process.
    1. Note that the Machine Learning Predict tool appears in the Machine Learning tool palette.

The integration tools are separate from the tools included in Alteryx Intelligence Suite.

tool Example

There is an example workflow for the Machine Learning Predict tool in the Alteryx Starter Kit for Intelligence Suite. To use the workflow:

  1. Go to the downloads page.
  2. Then go to Starter Kits > Alteryx Starter Kit for Intelligence Suite
  3. Download and install the Alteryx Starter Kit for Intelligence Suite EXE file.
  4. Once installed, open Designer and go to Help > Sample Workflows > Intelligence Suite Starter Kit > (your language).
  5. Select "Part 2: Scoring New Data with an Alteryx Machine Learning Model" to open the example workflow.
    1. Note that you must first run the "Part 1" workflow for the Machine Learning Send tool to build the model. 

Configure the Tool

Connect to Alteryx Machine Learning

If you want to run the Machine Learning Predict tool on Server, follow these additional steps. If not, you can follow these steps:

  1. First, you need to create an access token if you don’t already have one. Follow these steps to create one from the platform home page.
  2. Next, in the tool configuration panel, enter the URL that you use to access your Alteryx Machine Learning account.
  3. Select Sign In.
  4. Add a new data source in the Connection Manager. The Connection Manager should appear after you enter the URL for the first time.
    1. To open the Connection Manager, go to File > Manage Connections.
  5. Select + Add Data Source.
  6. Enter a Data Source Name.
  7. Under Technology, select Trifacta Auth if not already selected.
  8. Select Save.
  9. Within your new data source page under Connections, select Access Token under Authentication Method.
  10. Check the Allow connection for SDK box.
  11. Under Credential, select your authorization credential.
  12. If you don’t have an authorization credential, select Create New Credential.
    1. Enter a Credential Name.
    2. Enter your Access Token from step 1.
  13. Select Create and Link.
  14. Select Link.
  15. Once you are successfully connected to your account, the tool displays an "Authentication Verified" message.
  16. To change credentials, select Change Credentials.

Modeling Selection

  1. Prepare your new data for model prediction.

    1. Make sure to remove the target variable from your dataset.

    2. Match column names and data types to the dataset used to train your model.

  2. From the Machine Learning Predict tool, select the model you want to apply to your new data by selecting your Project.

    1. The Project Details, Model Details, and Ranking Metrics are all filled in from the model you've built in Alteryx Machine Learning.

  3. Select Run to apply the model to your new data.

  4. The Machine Learning Predict tool creates new columns of data in the output, which contain predictions.

To avoid type warnings in your workflow, use a typing tool like the Select tool or Auto Field tool to change the column type. You can also manually fix type warnings.

Was This Page Helpful?

Running into problems or issues with your Alteryx product? Visit the Alteryx Community or contact support. Can't submit this form? Email us.