SANN - Results - Liftcharts Tab

Neural Networks

Select the Liftcharts tab of the SANN - Results dialog box to access the options described here. This tab is only available for classification models. For information on the options that are common to all tabs, see SANN - Results. Note that for all lift chart types, you can include cases in the Train, Test, and/or Validation subsets by selecting the appropriate check boxes in the Sample group box. For example, to view a gains chart of all categories for only the Validation subset, select the Gains chart option button, select All in the Category list, select the Validation check box in the Sample group, and click the Lift chart button. Only the results for cases in the validation sample will be plotted.

Type. The options in this group box are used to create lift charts and gains charts for the categories of the target variables and for the current model. Use these charts to evaluate and compare the utility of the model for predicting the different categories or classes for the categorical target variable. Select the option button that specifies the type of chart and the scaling for the chart you want to compute.

Gains chart. Select this option button to compute a gains chart. This chart shows the percentage of observations correctly classified into the chosen category when taking the top x percent of cases from the sorted (by classification probabilities) data file.

For example, this chart can show you that by taking the top 20 percent (shown on the x-axis) of cases classified into the respective category with the greatest certainty (maximum classification probability), you would correctly classify almost 80 percent of all cases (as shown on the vertical y-axis of the plot) belonging to that category in the population. In this plot, the baseline random classification (selection of cases) would yield a straight line (from the lower-left to the upper-right corner), which can serve as a comparison to gauge the utility of the respective models for classification.

Lift chart (response %). Select this option button to compute a lift chart where the vertical y-axis is scaled in terms of the percentage of all cases belonging to the respective category. As in the gains chart, the x-axis denotes the respective top x percent of cases from the sorted (by classification probabilities) data file.

Lift chart (lift value). Select this option button to compute a lift chart where the vertical y-axis is scaled in terms of the lift value, expressed as the multiple of the baseline random selection model.

For example, this chart can show you that by taking the top 20 percent (shown on the x-axis) of cases classified into the respective category with the greatest certainty (maximum classification probability), you would end up with a sample that has almost 4 times as many cases belonging to the respective category when compared to the baseline random selection (classification) model.

Category. Select the response category for which to compute the gains and/or lift charts. You can chose to produce lift charts for a single or all categories.

Cumulative. Select this check box to show in the chosen lift and gains charts the cumulative percentages, lift values, etc. Clear this check box to show the simple (noncumulative) values.

Lift chart. Click this button to create the chart as specified via the options above.

ROC. Use the options in this group box to create ROC (Receiver Operating Characteristic) curves and spreadsheets for the active neural networks and specified samples. Note that this group box is available only for binary classification problems.

ROC curve. Click this button to create a line plot of the ROC curve for any number of active neural networks in one window given the data samples. The same information can be generated in spreadsheet format using the ROC spreadsheet option.

ROC spreadsheet. Click this button to create two spreadsheets, one containing data (i.e., 1-specificity versus sensitivity) and the other containing estimates of the ROC areas and thresholds for each active neural network and specified data samples.