Abstract
A robustified residual autocorrelation is defined based onL 1-regression. Under very general conditions, the asymptotic distribution of the robust residual autocorrelation is obtained. A robustified portmanteau statistic is then constructed which can be used in checking the goodness-of-fit of AR(p) models when usingL 1-norm fitting. Empirical results show thatL 1-norm estimators and the proposed portmanteau statistic are robust against outliers, error distributions, and accuracy for a given finite sample.
Similar content being viewed by others
References
Martin, R. D., Robust estimation of autoregressive models (with discussion), inDirections in Time Series (eds. Brillinger, D.R., Tiao, G. C.), Hayward: Inst. Math. Statist. Pub., 1980, 228–262.
Basawa, I. V., Huggins, R. M., Staudte, R. G., Robust tests for time series with an application to first-order autoregressive processes,Biometrika, 1985, 72: 559.
Bustos, P. H., Yohai, V. J., Robust estimates for ARMA models,J. Amer. Statist. Assoc., 1986, 81: 155.
Martin, R. D., Yohai, V. J., Robustness in time series and estimating ARMA models, inHandbook of Statistics, 4 (eds. Brillinger, D. R., Krishnaiah, P. R.), New York: Elsevier, 1985.
Birkes, D., Dodge, Y.,Alternative Methods of Regression, New York: Wiley, 1993, 191–192.
Cade, B. S., Richards, J. D., Permutation tests for least absolute deviation regression,Biometrics, 1996, 52: 886.
An, H. Z., Chen, Z. G., On convergence of LAD estimates in autoregression with infinite variance,J. Multivariate Anal., 1982, 12: 335.
Dunsmuir, W. T., Spencer, N. M., Strong consistency and asymptotic normalityL 1 estimates of the autoregressive movingaverage models,J. Time Series Anal., 1991, 12: 95.
Rao, C. R., Methodology based on theL 1-norm, in statistical inference,Sankhya A, 1988, 50: 289.
Chen, X., Asymptotical normality of least absolute deviations estimation in linear models,Science in China (in Chinese), Ser. A, 1990, 5: 449.
De Angelis, D., Hall, P., Young, G. A., Analytical and bootstrap approximations to estimator distributions inL 1 regression,J. Amer. Statist. Assoc., 1993, 88: 1310.
Box, G. E. P., Pierce, D. A., Distribution of the residual autocorrelations in aoturegressive integrated moving average time series models,J. Amer. Statist. Assoc., 1970, 65: 1509.
Koul, H. L., Zhu, Z., Bahadur-Kiefer representations for GM-estimators in autoregression models,Stochastic Processes and Their Applications, 1995, 57: 167.
Koenker, R., Bassett, G., Tests of linear hypotheses andl 1 estimation,Econometrica, 1982, 50: 1577.
Li, W.K., A goodness-of-fit test in robust time series modelling,Biometrika., 1988, 75: 355.
Martin, R. D., Samarov, A., Vandaele, W., Robust methods for ARIMA models, inApplied Time Series Analysis of Economic Data (eds. Zellner, A.), Washington D. C.: US Bureau of Census, 1983, 153–169.
Koenker, R., Zhao, Q., Conditional quantile estimation and inference for ARCH models,Econometric Theory, 1996, 12: 793.
Ruppert, D., Carroll, R. J., Trimmed least square estimation in the linear model,J. Amer Statist Assoc., 1980, 75: 828.
Stout, W. F.,Almost Sure Convergence, New York: Wiley, 1974.
Bickel, P. J., One-step Huber estimates in linear models,J. Amer. Statist. Assoc., 1975, 70: 428.
Author information
Authors and Affiliations
Additional information
Project supported by the Foundation of State Educational Commission and a research grant from the Doctoral Program Foundation of China (#97000139).
Rights and permissions
About this article
Cite this article
Jiang, J., van Hui, Y. & Zheng, Z. Robust goodness-of-fit tests for AR(p) models based onL 1-norm fitting. Sci. China Ser. A-Math. 42, 337–346 (1999). https://doi.org/10.1007/BF02874252
Received:
Issue Date:
DOI: https://doi.org/10.1007/BF02874252