Kernel Methods for Discrete-Time Linear Equations

摘要

Methods from learning theory are used in the state space of linear dynamical systems in order to estimate the system matrices and some relevant quantities such as a the topological entropy. The approach is illustrated via a series of numerical examples.