Application Of BP Network In Choas Time Series Forecast

Choas time series analysis which includes very abundant and profound meaning haswidely applications. In this dissertation, Choas time series analysis theory, method and BPnetwork's application to Choas Time Series forecast are studied. Its validity and easibility areproved by experiments use Matlab program.The main contents can be summed up into the following:1) Some forecast model are compared, and educed a conclusion about BP network'sadvantage when it was applied to Choas Time Series forecast domain.2) Phase space reconstruction theory: Phase space reconstruction is a valid method whichcan extract the intrinsic information of time series. The selection methods of time delay,embedding dimension of phase space reconstruction and Lyapunov exponent computation arethoroughly studied. Concise algorithms are given.3) Neuro-network theory is applied to Chaos Time Series forecast domain, and educefunction relation after BP network's reconstruction, addition with the realize procedure.Through the trained network's forecast, we proved this method was effective, and applicationsexample was given.4) Parameter enactment method of BP network is brought out, we also discuss the method ofparameter's optimize.