How to: Select a Predictable Column for an Accuracy Chart

Before you can create an accuracy chart, you must select a predictable column. All models that you include in the chart must use the same predictable column. You can also specify a target value for the predictable column, or leave the Predict Value field blank.

Whether you include or exclude a value for the predictable column of the model determines which of two different types of charts Analysis Services creates.

  • If you include a Predict Value, you will see a chart that shows how much lift the model provides for the target value.

  • If you do not include a Predict Value, you will see a chart that shows how accurate the model is for all predictable values.

For more information about the difference between the two types of charts, see Lift Chart (Analysis Services - Data Mining).

To select the predictable mining columns

  1. Double-click the mining structure that you want to chart to open it in Data Mining Designer.

  2. Select the Mining Accuracy Chart tab.

  3. Select the Input Selection tab.

  4. On the Input Selection tab, under Predictable Column Name, click the row and select a predictable column for each model.

    The mining model columns that are listed in the Predictable Column Name box are restricted to columns that have the usage type set to Predict or Predict Only. The mining structure column on which this mining model column is based must also have a content type of Discrete or Discretized.

  5. If you want to determine the lift for a model, select a value for the model from the Predict Value list.

See Also

Other Resources