Target Variable Step

Use the options in the Target variable step to define the quality criterion, i.e., what constitutes acceptable and unacceptable cases in the data. This step is only required if there are more than ten input variables in the analysis.

This is a vital step in the model building process. The inclusion of an input in a network adds another dimension to the space in which the data cases reside. The more inputs a neural network has, the more data points we need to train the network effectively. The goal of this step is to identify those inputs that carry a small amount of information on the problem and eliminate them from the analysis. Although this may lead to some loss of information in the data, it may, alternatively, considerably reduce the dimensionality, which can significantly increase the performance of the neural network.

The Important variables step is performed separately for each target variable, and it contains two tabs: Important variables and Annotations.