The Statistics Neural Networks Module allows you to identify complex relationships between data thanks to non-linear modeling procedures, called neural networks (as they are inspired by biological nervous systems, such as neurons in the brain). The models included in Neural Network allow you to:

​

• Define the conditions of learning of the network

• Check the rules of break of the network

• check the architecture of the network

• Choose the network architecture automatically

 

With Statistics Neural Networks it is possible to develop accurate predictive models, whose main characteristics are:

 

• Set templates Data mining of hidden relationships, using the Multilayer Perceptron (MLP) or Radial Basis Function (RBF) procedure

• Check the whole process by specifying the variables

• Integrate other statistical techniques or procedures to improve results

 

Data mining of hidden relationships

 

• Choose MLP or RBF algorithms to match the relationships implied by your data. The MLP procedure allows you to find the more complex relationships while the RBF procedure is faster.

• Benefit from feed-forward architectures, which move data in only one direction, from input nodes through the hidden layer or layers of nodes to output nodes.

• Take advantage of algorithms that operate on a training dataset and then apply those skills to the entire dataset and new data.

 

Check the process

 

• Specify dependent variables, which can be scale, categorical, or a combination of both.

• Modify the procedures by choosing how to partition the dataset, which architectures to use and which computational resources to apply to the analysis.

• Choose whether to view results in tables or graphs, save optional temporary variables in the active dataset, or export models to XML format for future data classification.

 

Complement other statistical techniques or procedures

 

• Confirm neural network results with traditional statistical techniques using Statistics Base.

• Integrate the other statistical procedures to obtain clearer information in different areas. For example, in market research, it is possible to create customer profiles and identify their preferences.