Development the system analysis methodology for modeling and assessment of financial risks.
The method of risk assessment was proposed for analysis the financial processes using the principles of systems analysis. Developed original approach includes following steps: (1) the assessment of the financial state, efficiency and business activity the analyzed subject; (2) the analysis of input data for identification structure and parameters uncertainties; (3) building the mathematical models for assessment the current and forecasting future states for risks calculation of the subject under study; (4) creating the set of candidates-scenarios of process and choosing the optimum decision. New methods for approximating loss function were developed to solve the described task. Those methods take into account nonlinear properties of the input data for the process. The methods for assessing possible losses based on the VaR methodology and its modifications were realized as a software application using contemporary environments SAS Base, SAS IML and SAS / SQL. The information technology uses contemporary principles of building cross-platform information system that enables to implement the proposed system analysis method, implemented methods, approaches and mathematical models in other information systems. Selected statistics for performing computational experiments, which aimed for analyzing the types of probability distributions, identification patterns of selected processes and effects caused by domestic financial institutions and features of random external influences. The method for assembling the forecasting results of losses obtained by various modeling techniques. The method is based on the application of a weighted combination of assessments and optimization approach that provides optimal computing weights. The proposed methods were applied for the analysis of real financial statistical data. The positive results of the numerical experiment were obtained. The estimation of possible losses corresponds to the actual financial data.