A new characterization of chaos from a time series

In the reconstruction scheme, the global fitting is a basis for the approach to the time evolution of dynamic systems directly from time series. A new theory of dynamic characterization is present in the aim of this work. The least squares method determines the predictors in the Algebraic Computation environment. The program for diagnosing of time series run in a Maple environment. The computational routine determines a new quantifier of chaos. A test for theory and computational tools in periodic, chaotic and random systems is in the scope of this paper. An application of the method in a real-world time series gives a satisfactory result.