3.4 Basic Examples

Each neural network problem follows the same basic steps: (1) load the data set; (2) initialize the network; (3) train the network; and (4) validate the model. Here the basic steps common to all neural network problems are elucidated with two examples: a classification problem and a function approximation problem. The options that alter or go beyond the basic steps are particular to the neural network model and so are discussed in later chapters. A detailed discussion of the meaning and manipulation of NeuralFit output can be found in the chapters that discuss individual network types.