Uncertain nonlinear time series analysis with applications to motion analysis and epidemic spreading

Uncertain nonlinear time series analysis is a set of statistical techniques that use uncertainty theory to predict future values via nonlinear dynamics based on the previous observations. By assuming that the disturbance term is an uncertain variable, an uncertain nonlinear time series model is derived in this paper. In addition, this paper presents a method to estimate unknown parameters in an uncertain nonlinear time series model. Finally, some real examples (motion analysis and epidemic spreading) are provided to illustrate uncertain nonlinear time series analysis. As a result, it is shown that the uncertain nonlinear time series model may provide higher forecast accuracy than linear one.