Chaos time series which includes very abundant and profound meaning has widely applications. In this dissertation, Chaos time series analysis theory, method and Genetic neural network's application to chaos time series forecast are studied. Its validity and facility are proved by experiments use MATLAB program.Firstly,chaos time series are introduced and several forecast model are compared,and educed a conclusion about BP network's advantage when it was applied to chaos time series forecast domain.We expatiate the development of the artificial neural network technology and its status for non-linear forecast .Then the commonly used BP network model, structure, algorithm of improving, the basis of genetic algorithm, the steps to achieve, and a brief description of the neural network toolbox and genetic algorithm toolbox were given. The selection of the space dimension on the Phase space reconstruction and Lyapunov exponent estimates are discussed. Concise algorithms are given and parameter enactment method of BP network is brought out. and gives a simple algorithm and general approach of network parameters. Considering that the number of nerve cell in hidden layer, initial weight and unit's bias value are the most important factors to the forecasting's precision of ANN,genetic algorithm is used to choose a more reasonable frame of ANN. Genetic algorithm is good for deciding the proper fabric of net, and help the ANN to conquer its disfigurement. GA-BP algorithm makes use of the strongpoint of GA and BP algorithm,the results of the example shows that GA-BP algorithm is better than BP algorithm only. Neural-network theory is applied to Chaos Time Series forecast domain,and educe function relation after BP network's reconstruction,addition with the realize procedure. Through the trained network's forecast,we proved this method was effective,and applications example was given. |