Kaidanovych D. Modeling of economic systems and processes using counter-propagation neural networks

There are developed in the dissertation the conceptual and methodological principles of modeling of economic systems and processes based on counter-propagation neural networks. In particular, according to the proposed conception there are developed the methodological approach for construction of economic and mathematical models based on the identification of typical patterns of behavior of economic systems with the purpose of analysis and forecasting of their development, which fundamentally distinguishes this approach from alternative methods. There are constructed on the basis of the approach the models for predicting financial indexes and modeling the dynamic of stock rates within the portfolio of securities. There are also constructed the models of bankruptcy diagnosis within the developed methodological approach.