Create Algorithmic Traits
To create an algorithmic trait, go to Traits and follow the steps below:
- Click Create New Trait and select Algorithmic from the drop down menu.
- In the Basic Information section
- Name the trait.
- Select a data source.
- Choose a storage folder.
- Expand the Configuration pane and click Browse All Models . This opens a new window that lets you select the model you want to use with the trait.
- Select a model and click Add Selected Model to Trait .You cannot replace selected models for existing algorithmic traits. To use a different model, you must create a new algorithmic trait.Adding the model exposes the reach and accuracy settings.
- Select reach or accuracy as your goal and choose a value from the respective drop down menus. Click Save when done.
Configuration Settings for Algorithmic Traits
In Trait Builder, the Configuration section lets you associate an algorithmic model to a trait. To complete the algorithmic trait creation process, select a model and choose a reach or accuracy goal.
Configuration Fields and Settings
Select Model for Algorithmic Trait
Click the Update button to open the models window. From the window, select the algorithmic model that you want to use to create the trait.
Select Goal Accuracy
Select this option to create a trait based on accuracy. Accuracy is a scored value that indicates how close potential users are to your baseline. The accuracy scale ranges from 0 (least accurate) to 1 (most accurate).
Reach and Accuracy Data Columns
Located on the right, this section displays up to 21 rows of numeric data that displays the accuracy and reach values for your model.
Reach and Accuracy Slider
Located at the bottom of the graph, the slider lets you set a numeric value for your reach or accuracy goals. You can set the slider and then choose the reach or accuracy goal button to create a trait.