Zeitreihenzerlegung mittels des mehrkomponentigen Verallgemeinerten Berliner Verfahrens

Abstract

The subject of this thesis is the multi-component Generalized Berlin Method (mehrkomponentiges Verallgemeinertes Berliner Verfahren, VBV), a method oftime series analysis, with approximates a given time series through the approximation of its influence components based on differential splines.After a detailed introduction to VBV, including, in particular, differential splines, its efficiency as well as some of its applications will be discussed.Moreover, the investigation of the smoothing parameters is carried out by means of (generalized) cross-validation. Last but not least, the best linear unbiased prediction (BLUP) in case of VBV is analyzed.