SANN Overviews - Selecting the Input Variables

The number of input and output units is defined by the problem. The target variable is believed to depend on the inputs and so its choice is clear. Not so when it comes to selecting the inputs. There may be some uncertainty about precisely which inputs to use. It should be noted that using a sufficient number of right inputs is a matter of great importance in predictive modeling (i.e., relating a target variable to a set of inputs) and indeed all forms of statistical analysis. By including irrelevant inputs, for example, one may inadvertently damage the performance of  the neural network. On the contrary, a data set with an insufficient number of inputs may never be accurately modeled  by a neural network.