Parameter Estimation Methods for Fourier Regression Time Series Model

This paper present a multiple form of Fourier regression analysis.This model consist of one periodic response variable and severalindependent variables decompose into periodic componentsin order to have an uncorrelated predictors series. Ordinaryleast square and maximum likelihood methods of obtaining theparameters of the model was proved, the estimators are shownto be unbiased, variance of the estimators and variance of errorterm were derived and from the decomposed sum of square thecoecient of determination and adjusted coecient was derived.The steps of test of hypothesis are stated and the performanceof the error are checked based on Durbin Watson statistic tojustify the consistency and reliability of the estimation methods.