Article
Full entry |
PDF
(1.5 MB)
Feedback
PDF
(1.5 MB)
Feedback
Keywords:
correlation coefficients; Bartlett approximations; simulation studies
correlation coefficients; Bartlett approximations; simulation studies
Summary:
An explicit formula for the correlation coefficient in a two-dimensional AR(1) process is derived. Approximate critical values for the correlation coefficient between two one-dimensional AR(1) processes are tabulated. They are based on Bartlett’s approximation and on an asymptotic distribution derived by McGregor. The results are compared with critical values obtained from a simulation study.
References:
[1] Anděl J.: Modern trends in multivariate time-series analysis. Math. Operationsforsch. Statist., Ser Zbl 0379.62075
[2] Anděl J.: Statistische Analyse von Zeitreihe.
[3] Bartlett M. S.: Some aspects of the time–correlation problem in regard to tests of significance. J. Roy Zbl 0012.11401
[5] Goodman L. A., Grunfeld Y.: Some nonparametric tests for comovements between time series. J. Amer. Statist Zbl 0108.15602
[8] Haugh L. D.: Checking the independence of two covariance stationary time series: a univariate residual cross–correlation approach. J. Amer. Statist. Assoc. 71 (1976), 378–385 DOI 10.1080/01621459.1976.10480353 | MR 0418379 | Zbl 0337.62061
[9] McGregor J. R.: The approximate distribution of the correlation between two stationary linear Markov serie.
[10] McGregor J. R., Bielenstein U. M.: The approximate distribution of the correlation between two stationary linear Markov series II. Biometrika 52 (1965), 301–302 MR 0207165 | Zbl 0143.40601
[12] Nakamura A. O., Nakamura M., Orcutt G. H.: Testing for relationships between time series. J. Amer. Statist Zbl 0329.62070
Search
Partner of
