PREDICTION OF CHAOTIC TIME-SERIES WITH NOISE

In this paper, we propose algorithm of deterministic nonlinear predict ion, or a modified version of the method of analogues which was origin ally proposed by E.N. Lorenz (J. Atom. Sci., 26, 636-646, 1969), and a pply it to the artificial time series data produced from nonlinear dyn amical systems and further corrupted by superimposed observational noi se. The prediction performance of the present method are investigated by calculating. correlation coefficients, root mean square errors and signature errors and compared with the prediction algorithm of local l inear approximation method. As a result, it is shown that the predicti on performance of the proposed method are better than those of the loc al linear approximation especially in case that the amount of noise is large.