Systemic Approach to Forecasting

Abstract

Background . Further enhancement of forecasts quality for dynamics of financial and economic processes requires development of new techniques and approaches in the frames of modern concepts for constructing informational decision support systems (DSS). Objective . The main purpose of the study is as follows: to consider system analysis principles that are suitable for solving the problem of short-term forecasting; to develop effective data processing system that implements the system analysis principles selected in the frames of DSS; to analyze possible types of uncertainties that are encountered in model constructing and forecasts estimating, and to propose the methods for their description and taking into consideration. Methods . To develop DSS for forecasting financial and economic processes and estimation of financial risks the following system analysis principles were hired: hierarchical architecture, the possibilities for identification and processing possible uncertainties, alternatives computing, and tracking the computational procedures for all stages of data processing. The system developed provides possibilities for taking into consideration statistical and parametric uncertainties. The DSS proposed has a modular architecture that could be easily expanded with new functions like preliminary data processing, model parameters estimation, and procedures for computing forecasts and financial risks. Results . The main result of the study is systemic methodology of mathematical modeling financial and economic processes, that has been implemented in the frames of the DSS proposed. High quality of final results is achieved thanks to appropriate tracking of all computations using several sets of statistical quality criteria. An example is given for mathematical modeling, estimation and forecasting of financial risk. The results of estimation show that the systemic approach proposed has good perspectives for its practical use. Conclusions . Thus, we proposed a systemic approach to mathematical modeling and forecasting financial and economic processes as well as estimation of financial risk. The use of the approach provides possibilities for computing estimate forecasts of high quality using statistical data.