View Performance Charts

View the Training Distribution Chart

To demonstrate the Variable Distribution chart features:
  1. In the SAS Model Manager Project Tree, expand Tutorial3, expand Loan, and expand 2013.
  2. Select the Performance node to display the Performance charts.
    Performance Node Charts
  3. On the Variable Distribution tab, click the Select variable box and select MORTDUE. The training distribution data and charts display the data for the MORTDUE variable.
    Training Distribution Data
  4. Click on a bar to highlight the corresponding entry in the table. If necessary, move the scroll bar to locate the highlighted entry in the table.
  5. In the table, click the highlighted row and drag the mouse to select multiple rows. The associated bars in the chart are highlighted as you select the associated rows.
    Highlighted Rows in the Training Distribution Table and Chart

View the Characteristic and Stability Charts

To demonstrate the Characteristic and Stability chart features:
  1. Select the Characteristic and Stability tab.
  2. Select table entries to highlight the corresponding chart points.
    Characteristic and Stability Charts

View the Lift Chart

To demonstrate the Lift chart features:
Note: This chart is not displayed for prediction models with an interval target.
  1. Select the Lift tab.
    Lift Chart as a Line Plot
  2. To change the chart to show markers:
    1. Right-click the chart and select Graph Properties.
    2. In the Properties window, select Line from the left-side menu.
    3. Click the Markers tab, and select the Show Markers check box.
    4. Click Apply to save the changes, and then click OK to close the window.
    Lift Chart with Show Markers
  3. Move the pointer along one lift marker. You should see a pop-up message when the pointer is resting on a data point or is close to a data point.
    Pop-up message when the pointer is on a data point.
  4. To chart the cumulative captured response:
    1. Right-click the chart area.
    2. Select Data Options. The Data Options Dialog window appears.
    3. Click the Roles tab if it is not already displayed. Select Y from the Roles drop-down list.
    4. Select the variable cuCapturedResp from the Available Variables list box and click the right-arrow to move it to the Assigned Variables list box.
    5. Select the cuLift variable and click the left-arrow to remove it from the Assigned Variables list box. The only variable in the Assigned Variables box should be cuCapturedResp.
    6. Click OK.
      The Lift chart displaying the variable CuCapturedResp.

View the Gini (ROC and Trend) Charts

To view the ROC and Gini charts, select the GINI (ROC and Trend) tab.
Note: This chart is not displayed for prediction models with an interval target.
Gini (ROC and Trend) chart

View the Kolmogorov-Smirnov (KS) Chart

To view the KS chart from the SAS Model Manager user interface:
Note: This chart is not displayed for prediction models with an interval target.
  1. Select the KS tab.
    KS Chart
  2. Select a different time point from the drop-down list of the Time Interval field.

View the Mean Squared Error (MSE) Chart

To view the MSE chart from the SAS Model Manager user interface:
Note: This chart is displayed only for prediction models with an interval target.
  1. In the SAS Model Manager Project Tree, expand Tutorial3, expand HMEQ-Interval, and expand 2013.
  2. Select the Performance node to display the Performance charts.
  3. Select the MSE tab.
    MSE Chart