Application Of Fourier Series Analysis To Temperature Data

Abstract

This Paper seeks to model a periodic time series using Fourier Series Analysis Method and to use such model to forcast future values of such data. The mean monthly temperature of Uyo Metropolis consisting of 180 data points (1991 – 2006) are collected for the study. The parameter estimates of the Fourier series model are obtained by ordinary least squares method in multiple regression. The test of significance of the general model and parameters indicate that the model is statistically significant and the significant parameters provide a Fourier series model of the form: 26.82-1.163cosωt - 0.169 sinωt + 0.133cos2ωt +0.164sin2ωt - 0.116sin4ωt + 0.255et-1. The P – P plot is also used to test for the overall goodness-of-fit and it is found out that the model fits well to the data and can be used to forcast the future values of the data.

Keywords: Fourier Series Analysis, Periodic Time Series, Forcasting.

Global Journal of Mathematical Sciences Vol. 7 (1) 2008: pp. 5-14