Increase of data proceeding relibility on the basis of fuzzy model of random time processes identification

The task is formulated and methodical bases are developed to construct the systems of data intellectual analysis on the basis of identification by hybrid model for increase of random time processes forecast reliability. Approaches are proposed to constructing the new mechanisms for optimizing data processing on the basis of fuzzy models on the basis of identifying the nonstationary properties of random time processes and synthesis of statistical and dynamic dentification models. The principles, methods and algorithms of identification based on combining the capabilities of fuzzy sets, fuzzy logic and a neural network are developed.The reliability of the developed technique is confirmed by results of experimental researches.