Under the Inputs and Outputs text boxes are four numeric values known as the Neural Network Parameters.  These parameters govern how the classification algorithm operates:

 

Number of clusters: the number of clusters, or categories, into which you want the data set divided.

 

Learning rate: a value between 0 and 1 that affects the rate at which the network learns.  The larger the learning rate, the faster the network will converge.  Be advised that oscillation and non-convergence can occur if the learning rate is set too high.

 

Epochs: the number of complete passes through the neural network of the entire set of training patterns.  Increasing this number is recommended when you have a large array of data.  Increasing the number of epochs can improve accuracy, but will slow down the classification.

 

Initial weights: By  increasing  initial  weights  you  increase  the weight of the initial condition / position of values of neurons in the network. By setting a low initial weight, initial positions of neurons will be almost random.

 
 

In most cases, the default values are acceptable for Learning rate, Epochs, and Initial weights.