Artificial Neural Networks: Theory and ApplicationsPrentice Hall, 1996 - 477 стор. "This is a comprehensive text on neural networks with a good balance between theory and applications. All of the important network architectures and learning algorithms are covered with a presentation of the underlying theory follow by typical applications. The book consists of seven parts: Introduction-Background and Biological Inspiration, Early Neural Networks and Developments, Multilayer Feedforward Neural Networks an Backpropagation, Dynamic Recurrent and Stochastic Neural Networks, Other Neural Network Architectures, Networks Based on Unsupervised Learning, and a concluding chapter on Neuro-fuzzt Systems, Soft Computing, Genetic Algorithms, and Neuro-Logic Networks. The latest developments in network architectures and learning algorithms are covered with extensive coverage given to dynamic recurrent networks and multilayer perceptrons and backpropogation learning"--Back cover. |
Зміст
Characteristics of Artificial Networks | 20 |
Review of Mathematical and Statistical Concepts | 37 |
Early Neural Network Architectures | 85 |
Авторські права | |
14 інших розділів не відображаються
Загальні терміни та фрази
activation functions ADALINE adaptive annealing applications attractor backpropagation behavior binary Boltzmann machine cells chapter classification components computed connections convergence defined Delta Rule derivative described dimension dynamics entropy equation estimate F2 layer feedback feedforward feedforward networks fuzzy logic fuzzy set given gradient descent Hebbian Hebbian learning hidden layer hidden units hidden-layer nodes Hopfield network illustrated in Figure initial input layer input pattern input vector learning algorithm learning process linear mapping methods minimize MLFF networks neocognitron network architecture neural network neuro-fuzzy neurons nonlinear operation optimization output nodes output units parameter pattern recognition perceptron performance plane predict problem random recurrent networks sample sigmoid function simulated simulated annealing single SOFM solution supervised learning target training examples training patterns training process training set variables vector quantization w₁ weight matrix weight values weight vector x₁ zero