6.2 BASIC CONCEPTS OF NEURAL NETWORKS: ARCHITECTURE
• Finally, how a network is trained to perform its desired task is another identifying model characteristic.
• With supervised learning, a sample training set is used to “teach” the network about its problem domain.
• This training set of exemplar cases (input and the desired output) is iteratively presented to the neural network.
• The output of the network in its present form is calculated and compared to the desired output.
• The learning algorithm is the training procedure that an ANN uses.
• The learning algorithm used determines hoe the neural interconnection weights are corrected due to differences in the actual and desired output for a member of the training set.
• Updating of the network’s interconnection weights continues until the training algorithm’s stopping criteria are met(e.g., all cases must be correctly classified within a certain tolerance level)