Twice-layered neural network self-organization for interpolation problems of artificial intelligence solution in the case of complete absence of information about small part of input variables

  • Authors:

  • A. G. Ivakhnenko;G. A. Ivakhnenko;E. A. Savchenko

  • Affiliations:

  • International Educational Centre of Informational Systems and Technologies of Nat. Ac. Sci. of Ukraine, Ukraine 01034 Kyiv. vul. Volodymirska 51/53 kv. 14;International Educational Centre of Informational Systems and Technologies of Nat. Ac. Sci. of Ukraine, Ukraine 01034 Kyiv. vul. Volodymirska 51/53 kv. 14;International Educational Centre of Informational Systems and Technologies of Nat. Ac. Sci. of Ukraine, Ukraine 01034 Kyiv. vul. Volodymirska 51/53 kv. 14

  • Venue:
  • Systems Analysis Modelling Simulation - Special issue: Self-organising modelling and simulation
  • Year:
  • 2003

Quantified Score

Hi-index 0.01

Visualization

Abstract

Method of diabetic home monitoring data for forecast is developed. Sloppy records with many breaks and even with the absent essential variables can be transformed into single-moment data sample that is used for twice-multilayered neural network self-organization. Filtration ability of neural network compensates inaccuracy of input data sample, and in the result rather accurate forecast of diabetic illness can be received at any moment of monitoring.