Kuznyetsova N. The informational technologies of the financial data Bayesian analysis

The thesis describes development of new models, algorithms and technologies for financial data analysis. Using the mathematical tool of Bayesian networks (discrete and hybrid) is proposed in the thesis, new integrated models on its base were built and the informational technologies for the financial data analysis were developed. The methodology of building and using Bayesian networks, which is characterized by the completeness of the modeling procedure and ensures the transparency and correctness of the network built for solving specific tasks is devised. It allows to logically structure the financial data handling. The integrated approach based on the Bayesian networks and another methods of the intellectual data analysis (decision trees and logistic regression) for financial data analysis is proposed and substantiated. The new integrated models based on the integrated approach are devised. These models were tested on solving the bank tasks and during financial data analysis. The informational technologies based on the integrated models as well as informational decision support system were developed and introduced in two banks and the exhibition company. The received scientific results are used in several educational courses at the NTUU "KPI". Keywords: Bayesian networks, informational technologies, integrated models, informational decision support system, integrated approach, Bayesian analysis.