NURBS Model For Chaotic Time Series

Time series has been studied extensively in the fields of statistics, signalprocessing, econometrics and mathematical finance, etc. The last few years havewitnessed the topic of stochastic versus chaotic time series analysis. One aspectof the difference between the stochastic method and the chaotic time seriesanalysis methods is that researchers admit more readily to looking at the databefore finally specifying a model. Most of the previous studies are based on thestochastic process. Nowadays, more and more researchers are paying theirattention to the chaotic process. Select the model depends on the data, andnonlinear time series analysis invariant attractor invariants can be well describedand quantified the behavior of chaotic systems dependent.Determine the target data's attribute, how to modeling the nonlinear timeseries is the key problem, the core of nonlinear time series modeling difficultproblems include: 1. describe non-linear; 2. achieve better approximation results;3. join the inherent nonlinear characteristics; 4. smooth solution to the problem;5. limited computing power Optimal algorithm; 6. the flexibility of the modelitself. This article is a new model for these problems.NURBS is a basic model commonly used in geometric modeling. Howeverit cannot be used in the dynamics field because of the absence of an importantfactor– time. Inspired by the segmented characteristics of the basis functions,this article proposes a time-based NURBS model that can be directly applied tothe analysis of dynamical systems. In additional, an instance of re-constructingmissing data from simulated workloads is illustrated in this article to prove theefficiency and accuracy of the new model.