An Introduction To Artificial Neural Network Modeling

an IntroDUCtIon to artIFICIal neUral netWork MoDelInG By George C. S. Wang Describes in simple terms how Artificial Neural Network model (ANN) works … ANN is related to regression in many ways, and shows how… ANN model is best suited for very short-term forecasting. A A rtificial Neural Network (ANN) modeling is simply a nonlinear regression modeling procedure except that an ANN model is usually much more complicated than other types of nonlinear models. The structure of an ANN model somewhat resembles the neural network of a human body, and the model is, therefore, called an Artificial Neural Network. The ANN name itself has already caused some confusion in forecast modeling. The terminology employed in its modeling process is very confusing; even many sophisticated modelers find it difficult to work with. For example, a simple ANN model has three layers like a neural network: An input layer, an output layer, and in between, a hidden layer with nodes. Layers, especially the hidden layer and nodes, are unfamiliar terms in regression modeling. However, once the model is specified and the modeling terminology is understood, ANN modeling is not as esoteric as it sounds. The first ...