scientific article; zbMATH DE number 3974713
[For the entire collection see Zbl 0591.00030.] Based on the Lotka-Volterra approach for systems analysis and the qualitative behavior of so-called exponential chains, a special binary code is introduced for these chains, the exponential code. From the correspondence between exponential chains and the code a transformation rule between chains and codes is deduced which can be used to construct a huge number of dynamical chains with similar qualitative properties to the exponential chains. These chains lead to possible models for time-series description. The aim is to approximate time-series by a sequence of local models from a dynamical chain concept with a high data-reduction efficiency. This leads to a multicriteria problem, which consists of decreasing the number of local models with small complexity (order) to represent a given time series with a mean square error which is as small as possible.
This page was built for publication:
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3740562)