Practical methods for treating serial correlations in experimental observations
Bureau International des Poids et Mesures (Retired), 92310 Sèvres, France
Serial correlations can have a significant effect on the statistical characterization of measured quantity values. Several methods for revealing the presence and the nature of correlations are presented in this paper; they include the Allan variance, spectral analysis and autocorrelation functions. A summary description of wavelet methods is also included. The relative strengths and weaknesses of the methods are considered. Emphasis is put on practical applications. An attempt is made to give an intuitive interpretation to each method.
© EDP Sciences, Springer-Verlag, 2009