Clustering Network

A cluster network consists of a number of class-labeled exemplar vectors (each represented by a radial neuron). The vectors are assigned centers by clustering algorithms such as K-Means, and then labeled using nearby cases. After labeling, the centers positions can be fine-tuned using Learned Vector Quantization.

General

Element Name Description
Detail of computed results reported Detail of computed results; if Minimal detail is requested, summary statistics for the trained network and a graph of the network architecture will be displayed; at the Comprehensive level of detail, sensitivity analysis and descriptive statistics spreadsheets will be displayed; the All results level will display the predictions spreadsheet.
Missing data Specifies the substitution method for missing data.
Apply memory limit Use this option to limit the maximum data size that can be processed; note that very large data problems may require significant memory and processing resources; modify the defaults only as needed.
Memory limit Use this option to set the maximum data size that can be processed.
Save/run network file By default (Don't save trained networks), the program will simply train the network, report the results, and then discard the trained network. Use the Save network file option to save the trained network in a specific file for future application to other data; use the Run network file option to apply a previously saved network to new data.
Network file name Specifies the name of the network file to save or run; this option is not applicable if the Save/run network file option was set to Don't save trained networks.

Units

Element Name Description
Normalize Normalize numeric input variables and codebook vectors.
Number of cluster units Specifies the number of cluster units for cluster neural network.

Centers

Element Name Description
Assign centers Specifies the algorithm to be used in assigning centers. The options are: random sampling, K-Means, and random values.
Labels centers Specifies an algorithm to assign class labels to the cluster network.
K value Specifies the K factors used to recover the neuron class; only applicable if KL nearest cases are selected to Label centers. The K nearest training cases to the neuron are located. The most common class label among these K is assigned to the neuron, provided that at least L are in common. If there are less than L in common, or there is a tie, the neuron is unlabeled.
L value Specifies the L factors used to recover the neuron class; only applicable if KL nearest cases are selected to Label centers. The K nearest training cases to the neuron are located. The most common class label among these K is assigned to the neuron, provided that at least L are in common. If there are less than L in common, or there is a tie, the neuron is unlabeled.
Voronoi value Specifies a stated proportion of the assigned units; only applicable if Voronoi neighbors are selected to Label centers. If less than the given proportion are in the most common class, a blank label is applied, indicating unknown.

LVQ

Element Name Description
Learned vector quantization Enable learned vector quantization (LVQ) training.
Phase one algorithm Specifies the algorithm variant to be used for Learned vector quantization; not applicable if Learned vector quantization is not selected.
Epochs Specifies the number of epochs over which the algorithm will run. On each epoch, the entire training set is fed through the network, and used to adjust the network weights and thresholds.
Learning rate 1 Specifies a learning rate start value.
Learning rate 2 Specifies a learning rate end value.
Epsilon Specifies the Epsilon value here. In LVQ 2.1. and LVQ 3, the definition of approximately the same distance is controlled by this parameter. Epsilon is typically between 0 and 1, usually less than 0.5.
Beta Specifies the Beta value here. In LVQ 3, when both nearest exemplars are of the same class as the training case, they are both moved towards the training case. This movement is more subtle than when mismatched exemplars are found - the usual learning rate is multiplied by beta, which is greater than 0 but significantly less than one.

Pruning

Element Name Description
Prunes inputs Specifies sensitivity-analysis based pruning after training.
Prunes ratio Specifies the sensitivity pruning ratio threshold; only applicable if Prune inputs was selected.

Classification

Element Name Description
Classification threshold Specifies whether to use a threshold.
Accept value Specifies the classification threshold to use; only applicable if a Classification threshold was requested.
Threshold K Specifies the K nearest neighbor control factors. By default K is 1, corresponding to the standard Kohonen winner takes all algorithm.
Threshold L Specifies the L nearest neighbor control factors. By default L is 0, corresponding to the standard Kohonen winner takes all algorithm.

Predicted values

Element Name Description
Subset to generate results Specifies the subset of observations to be used to compute predicted and residual values; only applicable if Comprehensive output or All results is selected as the Detail of computed results reported.

Deployment

Deployment is available if the Statistica installation is licensed for this feature.

Element Name Description
Generates C/C++ code Generates C/C++ code for deployment of predictive model.
Generates SVB code Generates Statistica Visual Basic code for deployment of predictive model.
Generates PMML code Generates PMML (Predictive Models Markup Language) code for deployment of predictive model. This code can be used via the Rapid Deployment options to efficiently compute predictions for (score) large data sets.
Saves C/C++ code Save C/C++ code for deployment of predictive model.
File name for C/C code Specify the name and location of the file where to save the (C/C++) deployment code information.
Saves SVB code Save Statistica Visual Basic code for deployment of predictive model.
File name for SVB code Specify the name and location of the file where to save the (SVB/VB) deployment code information.
Saves PMML code Saves PMML (Predictive Models Markup Language) code for deployment of predictive model. This code can be used via the Rapid Deployment options to efficiently compute predictions for (score) large data sets.
File name for PMML (XML) code Specify the name and location of the file where to save the (PMML/XML) deployment code information.