SANNAutomaticNetworkSearch Members

These members are related to Statistica Automated Network Search Dialog.

Functions

Name Description
Cancel
CheckNumberOfHiddenUnitsMLP Calling this function may re-set the number of hidden units for an RBF network provided that the initial settings were incompatible with the nature of the dataset. This function is used only in the automation mode.
CheckNumberOfHiddenUnitsRBF Calling this function may re-set the number of hidden units for an RBF network provided that the initial settings were incompatible with the nature of the dataset. This function is used only in the automation mode.
GoToResults Call this option to go to the Results dialog without creating any networks. To do so you must have at least one active network in your current analysis.
OK Return value: Integer.
SaveNetworksC Call this function to save the active networks in C language.
SaveNetworksCSharp Call this function to save the active networks as C# Programs.
SaveNetworksJava Call this function to save the active networks as Java Programs.
SaveNetworksJavaMapReduce Call this function to save the active networks as Java MapReduce Programs.
SaveNetworksPMML Call this function to save the active networks in PMML format.
SaveNetworksSAS Call this function to save the active networks as SAS Programs.
SaveNetworksSQL Call this function to save the active networks as SQL Programs.
SaveNetworksTeradata Call this function to save the active networks as Teradata C Programs.
SaveNetworksUDF Call this function to save the active networks as SQL UDF Programs.

Properties

Name Description
Application Returns application object. Return value: Object. This property is read only.
ExpHiddenActivationFunction Select this option to include the exponential activation function for the hidden units. MLP networks only. Return/assignment value: Integer.
ExpOutputActivationFunction Select this option to include the exponential activation function for the output units. MLP networks only. Return/assignment value: Integer.
IdentityHiddenActivationFunction Select this option to include the identity activation function for the hidden units. MLP networks only. Return/assignment value: Integer.
IdentityOutputActivationFunction Select this option to include the identity activation function for the output units (recommended for MLP networks). This is the only available option for RBF with sum of squares error function. Return/assignment value: Integer.
IterationDialogVisible Select this option to make the iteration dialog visible in automation mode. This function is not associated with any dialog options. Assignment value: Integer.
MaximumHiddenWeightDecay Specifies the maximum value of the input-hidden weight decay. MLP networks only. Return/assignment value: Double.
MaximumNumberOfHiddenUnits Specifies the maximum complexity of the Multilayer Perceptron (MLP) neural networks to be trained. Return/assignment value: Integer.
MaximumNumberOfHiddenUnitsRBF Specifies the maximum complexity of the Radial Basis Function (RBF) neural networks to be trained. Return/assignment value: Integer.
MaximumOutputWeightDecay Specifies the maximum value of the hidden-output weight decay. Return/assignment value: Double.
MinimumHiddenWeightDecay Specifies the minimum value of the input-hidden weight decay. MLP networks only. Return/assignment value: Double.
MinimumNumberOfHiddenUnits Specifies the minimum complexity of the Multilayer Perceptron neural networks to be trained. You can specify the number of hidden units exactly by setting the minimum equal to the maximum (see MaximumNumberOfHiddenUnits). Return/assignment value: Integer.
MinimumNumberOfHiddenUnitsRBF Specifies the minimum complexity of the Multilayer Perceptron neural networks to be trained. You can specify the number of hidden units exactly by setting the minimum equal to the maximum (see MaximumNumberOfHiddenUnitsRBF). Return/assignment value: Integer.
MinimumOutputWeightDecay Specifies the minimum value of the hidden-output weight decay. Return/assignment value: Double.
ModelListView Specifies the list of active models. Return/assignment value: String.
Name Return value: String. This property is read only.
NumberOfNetworksToRetain Specify how many of the networks trained by the Automated Network Search (ANS) should be retained. Return/assignment value: Integer.
NumberOfNetworksToTrain Specifies the number of networks to be trained by the Automated Network Search (ANS). Return/assignment value: Integer.
NumberOfTrainingCyclesANS Use this option to specify the number of training iterations. Assignment value: Integer.
Parent Returns the parent of the object. Return value: Object. This property is read only.
PrintDataStat Use this option to generate a statistics spreadsheet for the analysis variables. Return value: Object. This property is read only.
ResultsCodes Assignment value: Variant.
ResultsOption Assignment value: Integer.
ResultsOutputFields Assignment value: Variant.
ResultsSaveFileName Assignment value: String.
ResultsSelection Assignment value: Variant.
ResultsSelectionOption Assignment value: Variant.
ResultsValues Assignment value: Variant.
ResultsVariables Assignment value: Variant.
SeedValueForNetworkInitialization Use this option to specify the value of the seed for the random number generator used to initialize the network weights. Return/assignment value: Integer.
SigmoidHiddenActivationFunction Select this option to include the sigmoid activation function for the hidden units. MLP networks only. Return/assignment value: Integer.
SigmoidOutputActivationFunction Select this option to include the sigmoid activation function for the output units. MLP networks only. Return/assignment value: Integer.
SineHiddenActivationFunction Select this option to include the sine activation function for the hidden units. MLP networks only. Return/assignment value: Integer.
SineOutputActivationFunction Select this option to include the sine activation function for the output units. MLP networks only. Return/assignment value: Integer.
SummaryOfNetworks Print summary spreadsheet of active networks. Return value: Object. This property is read only.
TanhHiddenActivationFunction Select this option to include the tangent hyperbolic activation function for the hidden units (recommended). MLP networks only. Return/assignment value: Integer.
TanhOutputActivationFunction Select this option to include the tangent hyperbolic activation function for the output units. MLP networks only. Return/assignment value: Integer.
UseEntropyError Specify whether the Automated Network Search (ANS) should use cross entropy error function for training neural networks (classification problems only). Return/assignment value: Integer.
UseFixedSeedForNetworkInitialization Select this option to use a specified seed value for the random number generator used to initialize the network weights. Return/assignment value: Integer.
UseHiddenWeightDecay Select this option to use weight decay for the input-hidden weights. MLP networks only. Return/assignment value: Integer.
UseMLPNetwork Specify whether the Automated Network Search (ANS) should create Multilayer Perceptron (MLP) networks. Return/assignment value: Integer.
UseOutputWeightDecay Select this option to use weight decay for the hidden-output weights. Return/assignment value: Integer.
UseRBFNetwork Specify whether the Automated Network Search (ANS) should create Radial Basis Functions (RBF) networks. Return/assignment value: Integer.
UseSOSError Specify whether the Automated Network Search (ANS) should use sum of squares error function for training neural networks. Return/assignment value: Integer.

Events

Name Description
OnTrainingProgress Fired periodically to notify clients of training progress.