Modeling Dynamic Systems by a State Space Representation using Information Theory
Dynamic Systems, Theoretical Information Learning, Information States.
Information Theory (IT) is a field of mathematics that studies quantifying the information. Recently several studies with Theoretical information Learning (ITL) has been successful as a new approach to non-supervised machine learning, most often involving only static data. In the area of dynamical systems, the representation of the state space is a modern approach that works the concept of state variables and is defined in the time domain. This paper proposes the incorporation of descriptors of information theory in modeling of dynamic systems in the form of state space. This proposal allows for modeling the dynamic behavior of the system based only on the intrinsic information to it. This proposal presents a new concept we call information states.