configure dataset

The configure dataset command configures a dataset and defines its features. In the notebook, you will be guided through a template form that will prompt you for various inputs, such as the name of the data source, each feature name, and type and definition parameters, such as the source data sink for that feature.

Syntax

configure dataset

Parameters

1. To create a new dataset, select New from the Dataset drop-down list, and specify a name for the dataset.

   Or:

   To change an existing dataset, select the name of the dataset from the Dataset drop-down list.

2. To create a new feature, select Create New from the Feature drop-down list, and specify a name for the feature.

   Or:

   To change an existing feature, select the name of the feature from the Feature drop-down list.

3. Select the feature type.

4. Specify a name for the feature and optionally a description.

5. For basis features:

   1. Select the data sink from the drop-down list.

   2. Select the required feature.

6. For contextual features:

   1. If the feature will use an expression, select the Expression checkbox.

      This checkbox is only enabled if you installed an Aizen license that allows expressions.

      • If you want to use one of Aizen's built-in expressions, select them from the Built-in Expressions drop-down menu.

      • If you want to specify your own expression, leave the Built-in Expressions field blank, or select None and specify the required expression in the expression field.

   2. For aggregate features:

      1. Make sure to clear the Expression checkbox.

      2. Select the input Datasink.

      3. Specify the source column from the data sink.

      4. Select the aggregate function.

      5. Select the timestamp column from the drop-down list.

      6. As needed, specify the appropriate missing strategy to handle missing data for the input feature.

      7. Specify the window start and end values to compute the windowed aggregate values.

      8. Specify the required join key and the feature to map to that key.

7. Select if the input feature will be treated as a label or output feature.

8. Select the Materialize checkbox if the feature needs to be materialized and stored in the dataset. If you leave the materialize checkbox cleared, then the feature is only created in memory during dataset creation and used in the computation of other features.

9. Click on Add Feature to add the feature to the dataset configuration.

10. Repeat the previous steps to define all the dataset features.

11. Click the Save Configuration button to start saving the dataset configuration.

12. Click OK to persist the configuration.

Example