160 likes | 176 Views
Explore the selective properties of the Generalized Method of Data Handling (GMDH) criteria for inductive modeling by Volodymyr Stepashko. The book delves into the theory of Noise-Immunity Modeling as the foundation of GMDH, emphasizing the method of critical variances. Learn how GMDH offers an inductive approach for crafting top-notch forecasting models and conducting extrapolations to aid in decision-making. Discover the significance of external criteria in accounting for uncertainties automatically within model structures.
E N D
Selective Properties of the GMDH Criteria for Inductive Modeling Volodymyr Stepashko Kyiv, IRTC ITS Ukraine
General Problem Statement F– set of model structures С– criterion of a model quality Structure of a model: Estimation of parameters: Q – criterion of the quality of model parameters estimation
2. Basic Principles of GMDH as an Inductive Method Construction of optimal models for prediction and extrapolation with the purpose of control and decision making
Conclusions • Theory of the Noise-Immunity Modeling is the basis of the GMDH theory • Method of critical variances is the basic analytical tool of the theory • GMDH is an inductive method for construction of optimal forecasting models • External criteria provide implicit (automatic) account of uncertainties