Lee, Andy H.; Li, W. K.; Hui, Yer Van On the empirical influence function of the portmanteau statistic in AR(1) process. (English) Zbl 0861.62061 J. Jpn. Stat. Soc. 26, No. 1, 83-90 (1996). Summary: We study the effect on residual autocorrelations in the presence of atypical observations for the AR(1) process. Based on the empirical influence function of the portmanteau statistic, a measure of influence is derived. A confirmatory procedure coupling robust estimation with Monte Carlo tests is then formulated to assess the magnitude of the sample statistics. An example illustrates application of the methodology. MSC: 62M10 Time series, auto-correlation, regression, etc. in statistics (GARCH) 62F35 Robustness and adaptive procedures (parametric inference) Keywords:diagnostics; outliers; residual autocorrelations; AR(1) process; empirical influence function; portmanteau statistic; measure of influence; robust estimation; Monte Carlo test PDFBibTeX XMLCite \textit{A. H. Lee} et al., J. Jpn. Stat. Soc. 26, No. 1, 83--90 (1996; Zbl 0861.62061) Full Text: DOI