A method to improve the precision of chaotic time series prediction by using a non-trajectory

Due to the error in the measured value of the initial state and thesensitive dependence on initial conditions of chaotic dynamicalsystems, the error of chaotic time series prediction increases withthe prediction step. This paper provides a method to improve theprediction precision by adjusting the predicted value in the courseof iteration according to the evolution information of smallintervals on the left and right sides of the predicted value. Theadjusted predicted result is a non-trajectory which can provide a better prediction performance than the usual result based on thetrajectory. Numerical simulations of two typical chaotic mapsdemonstrate its effectiveness. When the prediction step getsrelatively larger, the effect is more pronounced.