SANN - Results

Neural Networks

Click the Train button or Go to results button in either the SANN - Automated Network Search (ANS) dialog box or the SANN - Custom Neural Network dialog box to display the SANN - Results dialog box. In addition to the options described below, this dialog box can contain several tabs depending on the analysis type. These tabs include Predictions, Graphs, Details, Custom predictions, Lift charts (only for classification problems); Predictions (Kohonen), Graphs (Kohonen), and Kohonen Graph (only for cluster problems); and Time Series (only for time series problems).

Active neural networks. The grid in the Active neural networks group box provides a quick view of the networks you have created for modeling your data. Note that if you have not trained any networks or if you have not selected any active networks, this grid will be empty.

Model details. See network data grid section for a detailed description.

Select\Deselect active networks. Click this button to display the Model activation dialog box, where you can select the networks that you want to display in the Active neural networks grid. When you click OK in the Model activation dialog box, the Active neural networks grid will only display the selected networks. Note that when a network is deselected in the Model activation dialog box, it will not be displayed in this grid (i.e., it is inactive); however, it will still be available for future selection. To select unwanted networks for deletion, click the Delete networks button.

Delete networks. Click this button to display the Model deletion dialog box, where you can select networks to be completely removed from all results. Once you select a network and click the OK button in the Model deletion dialog box, you will be prompted to verify the deletion and then the network will be discarded. If you want to remove the selected networks from the Active neural networks grid while still maintaining them for future selection, use the Select\Deselect active networks option.

Build more models with CNN. Click this button to display the SANN - Custom Neural Network dialog box and use the options to build additional models.

Build more models with ANS. Click this button to display the SANN - Automated Network Search (ANS) dialog box and use the options to build additional models.

Summary. Click the Summary button to generate a spreadsheet containing the summary details listed in the Active neural networks grid box. Note that if the Active neural networks grid is empty, this button will not be available.

Data statistics. Click the Data statistics button to generate a spreadsheet containing the mean, standard deviation, minimum value, and maximum value for each continuous variable in the analysis. These data statistics will be broken down by each sample (training, testing, and validation) and also reported for the overall data set.

Save networks. Note that, according to which analysis was selected in the Startup Panel, this drop-down list may not contain all the commands listed below. Click this button to display a drop-down list containing the following commands:

PMML. Click this command to display the Save PMML file dialog box, which contains options to store the active networks for future use. Note that this dialog box will be displayed only when the Active neural networks grid contains networks. Stored PMML networks can be opened by clicking the Load network files button in the SANN - New Analysis/Deployment Startup Panel.

C/C++. Click this command to display the Save C file dialog box, which contains options to store the active networks for future use.

C#. Click this command to generate code as C#.

Java. Click this command to generate code in Java script.

SAS. Click this command to display the Save SAS file dialog box, which contains options to save deployment code for the created model as SAS code (a .sas file). See also, Rules for SAS Variable Names.

SQL stored procedure in C#. Click this command to generate code as a C# class intended for use in a SQL Server user defined function.

SQL User Defined Function in C#. Click this command to generate code as a C# class intended for use as a SQL Server user-defined function.

Teradata. Click this command to generate code as C Computer language function intended for use as a user-defined function in a Teradata querying environment.

Deployment to STATISTICA Enterprise. Click this command to deploy the results as an Analysis Configuration in STATISTICA Enterprise. Note that appropriately formatted data must be available in a STATISTICA Enterprise Data Configuration before the results can be deployed to an Analysis Configuration.

Cancel. Click the Cancel button to exit the SANN - Results dialog box and return to the previous dialog box. Any selections made will be ignored.

Options. Click the Options button to display the Options menu.

Samples. Use the options in this group box to select the samples to use when displaying results. The Missing check box is only enabled if the data contains cases where all the inputs are present but the targets are missing. Note that this option is never available for time series analysis; see data handling for time series analysis. Other sub samples are specified using the options on the Sampling tab of the SANN - Data selection dialog box. Note that some results (such as the Summary spreadsheet) provide information for all subsets regardless of the selections made here.

Train. Select the Train check box to include the cases assigned to the training sample when displaying graphs and spreadsheet.

Test. Select the Test check box to include the cases assigned to the test sample when displaying graphs and spreadsheet.

Validation. Select the Validation check box to include the cases assigned to the validation sample when displaying graphs and spreadsheet.

Missing. Select the Missing check box to generate certain results (e.g., predictions) for cases in which all the inputs are present but one or more output (target) values are missing. Note that this check box is only available when casewise deletion of missing data has been used. For times series analysis this option is never available.

All missing cases in SANN are grouped together to form a sample called the missing sample. It is used to train the hidden units of RBF networks and to make predictions for MLP or RBF networks. When a data case falls into the missing sample category, only the network outputs are available. In other words, there are no residuals since one or more targets is not available.