Improving the Accuracy of Identification of Non-Stationary Objects Based on the Regulation of Model Variables

Researched and developed mechanisms for optimizing the identification of random time series based on searching unknown knowledge, hidden properties, patterns, relationships, features non-stationary objects under condition limited a priori information, uncertainty, processes. A generalized model RTS use neural networks, neuro-fuzzy networks dynamic models, as well fuzzy logic algorithms is implemented. Instruments data reliability control are obtained statistical, dynamic, intelligent approaches suitable conditions transmission incomplete, heterogeneous, partially given information with large parametric uncertainty. The principles methods in environment generalization properties statistical modeling methods. Algorithms correlations, trends, interrelations regularities data, forming training, test sets solving problems recognition, classifying images micro-objects predicting power supply indicators synthesized. results research implemented form software package identifying objects, which ensures adaptability variables, high processing performance accuracy results.