The Annals of Statistics

Canonical correlation analysis and reduced rank regression in autoregressive models

T. W. Anderson

Full-text: Open access

Abstract

When the rank of the autoregression matrix is unrestricted, the maximum likelihood estimator under normality is the least squares estimator. When the rank is restricted, the maximum likelihood estimator is composed of the eigenvectors of the effect covariance matrix in the metric of the error covariance matrix corresponding to the largest eigenvalues [Anderson, T. W. (1951). Ann. Math. Statist. 22 327-351]. The asymptotic distribution of these two covariance matrices under normality is obtained and is used to derive the asymptotic distributions of the eigenvectors and eigenvalues under normality. These asymptotic distributions differ from the asymptotic distributions when the regressors are independent variables. The asymptotic distribution of the reduced rank regression is the asymptotic distribution of the least squares estimator with some restrictions; hence the covariance of the reduced rank regression is smaller than that of the least squares estimator. This result does not depend on normality.

Article information

Source
Ann. Statist., Volume 30, Number 4 (2002), 1134-1154.

Dates
First available in Project Euclid: 10 September 2002

Permanent link to this document
https://projecteuclid.org/euclid.aos/1031689020

Digital Object Identifier
doi:10.1214/aos/1031689020

Mathematical Reviews number (MathSciNet)
MR1926171

Zentralblatt MATH identifier
1029.62053

Subjects
Primary: 62H10: Distribution of statistics 62M10: Time series, auto-correlation, regression, etc. [See also 91B84]

Keywords
Canonical correlations and vectors eigenvalues and eigenvectors asymptotic distributions

Citation

Anderson, T. W. Canonical correlation analysis and reduced rank regression in autoregressive models. Ann. Statist. 30 (2002), no. 4, 1134--1154. doi:10.1214/aos/1031689020. https://projecteuclid.org/euclid.aos/1031689020


Export citation

References

  • AHN, S. K. and REINSEL, G. C. (1988). Nested reduced-rank autoregressive models for multiple time series. J. Amer. Statist. Assoc. 83 849-856.
  • ANDERSON, T. W. (1951). Estimating linear restrictions on regression coefficients for multivariate normal distributions. Ann. Math. Statist. 22 327-351. [Correction (1980) Ann. Statist. 8 1400.]
  • ANDERSON, T. W. (1971). The Statistical Analy sis of Time Series. Wiley, New York.
  • ANDERSON, T. W. (1984). An Introduction to Multivariate Statistical Analy sis, 2nd ed. Wiley, New York.
  • ANDERSON, T. W. (1999a). Asy mptotic theory for canonical correlation analysis. J. Multivariate Anal. 70 1-29.
  • ANDERSON, T. W. (1999b). Asy mptotic distribution of the reduced rank regression estimator under general conditions. Ann. Statist. 27 1141-1154.
  • BOX, G. E. P. and TIAO, G. C. (1977). A canonical analysis of multiple time series. Biometrika 64 355-365.
  • LÜTKEPOHL, H. (1993). Introduction to Multiple Time Series Analy sis, 2nd ed. Springer, New York.
  • REINSEL, G. C. (1997). Elements of Multivariate Time Series Analy sis, 2nd ed. Springer, New York.
  • REINSEL, G. C. and VELU, R. P. (1998). Multivariate Reduced Rank Regression. Springer, New York.
  • Ry AN, D. A. J., HUBERT, J. J., CARTER, E. M., SPRAGUE, J. B. and PARROTT, J. (1992). A reduced-rank multivariate regression approach to aquatic joint toxicity experiments. Biometrics 48 155-162.
  • SCHMIDLI, H. (1995). Reduced-Rank Regression: With Applications to Quantitative StructureActivity Relationships. Springer, Berlin.
  • STOICA, P. and VIBERG, M. (1996). Maximum likelihood parameter and rank estimation in reduced-rank multivariate linear regressions. IEEE Trans. Signal Process. 44 3069-3078.
  • TIAO, G. C. and TSAY, R. S. (1989). Model specification in multivariate time series (with discussion). J. Roy. Statist. Soc. Ser. B 51 157-213.
  • VELU, R. P., REINSEL, G. C. and WICHERN, D. W. (1986). Reduced rank models for multiple times series. Biometrika 73 109-118.
  • STANFORD, CALIFORNIA 94305-4065 E-MAIL: twa@stat.stanford.edu