ARTICLE
ABSTRACT
In this paper selected aspects concerning the use of power trend in the forecasting process were considered. Approximate methods for estimating trend parameters (equations (12)-(15)) were proposed. The methods yielded results similar to results given by least squares method (LSM). Formula (35) determining ex ante error of the forecast determined by the approximate methods and LSM for random element additive model was defined. Computer simulations were done – including three models of random element (additive, multiplicative, mixed) – with the aim of determining the range of usefulness of logarithmic transformation method, LSM and the approximate methods.
KEYWORDS
forecasting, estimation of power trend parameters, forecast ex ante error, computer simulations
REFERENCES
[1] Chow G.C., [1995], Ekonometria, Wydawnictwo Naukowe PWN, Warszawa.
[2] Dahlquist G., Bjorck A., [1983], Metody numeryczne, PWN, Warszawa.
[3] Jurkiewicz T., Plenikowska-Ślusarz T., [2001], Algorytmy optymalizacyjne w estymacji nieliniowych funkcji regresji, „Wiadomości Statystyczne” z. 10, s. 10-15.
[4] Kmenta J., [1990], Elements of Econometrics (second edition), Macmillan Publishing Co.
[5] Prognozowanie gospodarcze. Metody i zastosowania, [2001], pod redakcją Cieślak M. Wydawnictwo Naukowe PWN, Warszawa.
[6] Purczyński J., [2008], Wybrane aspekty prognozowania z wykorzystaniem trendu wykładniczego, „Przegląd Statystyczny” z. 1, s. 27-44.
[7] Zeliaś A., Pawełek B., Wanat S., [2004], Prognozowanie ekonomiczne. Teoria, przykłady, zadania, PWN, Warszawa.
