Demkivsky Y. The Information technologies for analysis and forecasting of nonstationary processes.

Українська версія

Thesis for the degree of Candidate of Sciences (CSc)

State registration number

0407U003759

Applicant for

Specialization

  • 05.13.06 - Інформаційні технології

28-09-2007

Specialized Academic Board

Д 26.002.03

Educational and Scientific Complex "Institute for Applied System Analysis" of National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute"r

Essay

The thesis is devoted to the problem of development of decision support systems at the modeling and forecasting of non-stationary processes, which are described by time series. Different approaches to the automation of selection of class and structure of the models, which describe time series, were investigated. Technology of allocation, modeling and forecasting of deterministic and stochastic trends was suggested. New models for description of stochastic trends and also conditions for selection of better model according to the multitude of statistic parameters were proposed in the thesis. Threshold model of stochastic volatility for modeling and forecasting of time-varying dispersion was offered. Technology of parameter estimation of the suggested model is processed; exact examples and conditions of effective application of suggested models are described. Architecture and prototype of DSS at the modeling and forecasting of non-stationary processes were developed.

Files

Similar theses