Method for computer-aided learning of autogenous neural network with temporally sequential conditions in finite time interval, involves connecting conditions with temporally sequential inputs and outputs of dynamic system

摘要

The method involves connecting conditions with temporally sequential inputs and outputs of a dynamic system. The conditions are connected with each other and are connected with the inputs and the outputs over convectors. The convector permits temporally forward and temporally backward arranged information flow for error back publicizing, with which information flow is considered in the autogenous neural network before beginning of the time interval or after end of the time interval by a retro causal network. Independent claims are included for the following: (1) a method for the prediction of the outputs of a dynamic system based on inputs of the dynamic system; and (2) a computer program product with a program code.