THRESHOLD VALUE FOR DIAGNOSIS OF CHAOTIC NATURE OF THE DATA OBTAINED IN NONLINEAR DYNAMIC ANALYSIS

In this paper surrogate data method of phase-randomized is proposed to identify the random or chaotic nature of the data obtained in dynamic analysis: The calculating results validate the phase-randomized method to be useful as it can increase the extent of accuracy of the results. And the calculating results show that threshold values of the random timeseries and nonlinear chaotic timeseries have marked difference.

THE NON-LINEAR CHAOTIC MODEL RECONSTRUCTIONFOR THE EXPERIMENTAL DATA OBTAINED FROMDIFFERENT DYNAMICSYSTEM

The non_linear chaotic model reconstruction is the major important quantitative index for describing accurate experimental data obtained in dynamic analysis. A lot of work has been done to distinguish chaos from randomness, to calulate fractral dimension and Lyapunov exponent, to reconstruct the state space and to fix the rank of model. In this paper, a new improved EAR method is presented in modelling and predicting chaotic timeseries, and a successful approach to fast estimation algorithms is proposed. Some illustrative experimental data examples from known chaotic systems are presented, emphasising the increase in predicting error with time. The calculating results tell us that the parameter identification method in this paper can effectively adjust the initial value towards the global limit value of the single peak target function nearby. Then the model paremeter can immediately be obtained by using the improved optimization method rapidly, and non_linear chaotic models can not provide long period superior predictions. Applications of this method are listed to real data from widely different areas.

Nonlinear and chaotic analysis of a financial complex system

In this paper,determination of the characteristics of futures market in China is presented by the method of the phase-randomized surrogate data.There is a significant difference in the obtained critical values when this method is used for random timeseries and for nonlinear chaotic timeseries.The singular value decomposition is used to reduce noise in the chaotic timeseries.The phase space of chaotic timeseries is decomposed into range space and null noise space.The original chaotic timeseries in range space is restructured.The method of strong disturbance based on the improved general constrained randomized method is further adopted to re-deternination.With the calculated results,an analysis on the trend of futures market of commodity is made in this paper.The results indicate that China＇s futures market of commodity is a complicated nonlinear system with obvious nonlinear chaotic characteristic.

GEKF,GUKF and GGPF based prediction of chaotic time-series with additive and multiplicative noises

On the assumption that random interruptions in the observation process are modelled by a sequence of independent Bernoulli random variables, this paper generalize the extended Kalman filtering （EKF）, the unscented Kalman filtering （UKF） and the Gaussian particle filtering （GPF） to the case in which there is a positive probability that the observation in each time consists of noise alone and does not contain the chaotic signal （These generalized novel algorithms are referred to as GEKF, GUKF and GGPF correspondingly in this paper）. Using weights and network output of neural networks to constitute state equation and observation equation for chaotic time-series prediction to obtain the linear system state transition equation with continuous update scheme in an online fashion, and the prediction results of chaotic time series represented by the predicted observation value, these proposed novel algorithms are applied to the prediction of Mackey-Glass time-series with additive and multiplicative noises. Simulation results prove that the GGPF provides a relatively better prediction performance in comparison with GEKF and GUKF.

THE INFLUENCE OF THE DIFFERENT DISTRIBUTEDPHASE-RANDOMIZED ON THE EXPERIMENTAL DATA OBTAINEd IN DYNAMIC ANALYSIS

In this paper the influence of the differently distributed phase-randontized to the data obtained in dynamic analysis for critical value is studied.The calculation results validate that the sufficient phase-randomized of the different distributed random numbers are less influential on the critical value . This offers the theoretical foundation of the feasibility and practicality of the phase-randomized method.

THE MATRIC ALGORITHM OF LYAPUNOV EXPONENT FOR THE EXPERIMENTAL DATA OBTAINED IN DYNAMIC ANALYSIS

The Lyapunov exponent is important quantitative index for describing chaotic attractors. Since Wolf put up the trajectory algorithm to Lyapunov exponent in 1985, how to calculate the Lyapunov exponent with accuracy has become a very important question. Based on the theoretical algorithm of Zuo Binwu, the matric algorithm of Lyapunov exponent is given, and the results with the results of Wolf's algorithm are compared. The calculating results validate that the matric algorithm has sufficient accuracy, and the relationship between the character of attractor and the value of Lyapunov exponent is studied in this paper. The corresponding conclusions are given in this paper.