Comparative Analysis of Methods for Prediction of Economic Parameter
Olga V. Dudnik · Journal of Automation and Information Sciences · 2001
Some modern directions of the statistical analysis and prediction of dynamic series are described. On examples of two real dynamic series, namely, for the Ukrainian hrivna rate relative to the US dollar and the Russian rouble for the second half of 1997, the operation of such prediction methods as the Brown method, the method of exponential smoothing with application of the Laguerre orthogonal polynomials, the adaptive Kalman filter is investigated. The regression model of a process was also constructed for these series, which is later tested for heteroscedasticity. The results of application of the noted above methods and comparison of their potentials are presented.