Transferencia de información en redes basada en modelado. García, P., Rosa, L., Chiza, J., Méndez, P., Otero, C., Fernández, Y., & Silva, A. Jornadas internacionales de investigación científica, pages 223-227. UTN, 2016.
Jornadas internacionales de investigación científica [pdf]Paper  abstract   bibtex   
This paper presents a strategy to estimate the conditional probability densities necessary for determining the cause-effect relationship between components of complex systems. The scheme estimates the information transfer between these components using predictions of the evolution of the subsystems. The method, do not require discretization of the state space necessary in the direct calculation of probabilities in continuous signals. The performance of the strategy is shown using data produced by numerical simulations and experimental data from chaotic systems. The results appear to be useful in the design of control strategies, in the case of systems formed by several parts.

Downloads: 0