A Comparison of the Pseudo-Additive Mixed Fourier series Approach and the SARIMA Methodology to the Modeling of Rainfall in Uyo
Anyanso Chinomso
Department of Mathematics/ Statistics
Faculty of Science, Rivers State University, Port Harcourt
ABSTRACT
Accurate rainfall forecasting is very important to the economic development of a country. It is not just important to the government but also to individuals, farmers and private companies. This paper focuses on comparing the performances of two approaches to seasonal time series analysis. These approaches are the pseudo-additive mixed Fourier series approach and the SARIMA approach. The pseudo-additive Fourier series approach decomposes a time series into the traditional components in a mixed model. This is suitable for a time series with very small or zero values like that in the data used, while the ARIMA model has significant advantages especially in short run forecasting saz [2011], The time series analysis methods were used to model the monthly rainfall of Uyo in Akwa Ibom state, Nigeria . The data were monthly value for ten[10] years .A comprehensive outline of both analysis methods are presented in this paper as well as the advantages each have after the other . The performances were evaluated based on three[3] statistics; mean absolute error [MAE], mean absolute percentage error[MAPE] and mean squared deviation [MSD], The result at the end showed that the SARIMA model has a smaller MAE, MAPE and MSD values .As such, it is the better model.
