دراسة تجريبية لاستعمال طرائق حصينة لتقدير الانموذجين SARFIMA (0, D, 0)‎s و SARFIMA (0, d, 0)‎ (0, D, 0)‎s

Abstract EN

The presence of outliers in any data affects the estimation of the parameters of these data.

Therefore, scientists have found invulnerable methods of estimation that exclude the effect of outliers on the data.

Among these methods is the repeated weighted least squares method and the use of different weight functions with it.

What we have done in this research is the use of this method in estimating the normal and seasonal difference parameters of the two models SARFIMA (0, d, 0) (0, D, 0) _S and SARFIMA (0, D, 0) _S using the (Gewek and Porter-Hudak) method based on the periodogram and spectrum function analysis.

It approximates the shape of the regression function, as well as using a periodogram adapted with different spectral windows