Robust Cumulant Estimation
One of the problems in the application of higher-order statistics (HOS) is that of the estimation of cumulants. The higher the order the larger tends to be the variance in the estimated cumulants and this problem is also enhanced by the limited number of samples used in applications. Naturally, the accuracy of the methods based on higher—order statistics depend on, among other things, the consistency of the estimates of the cumulants. Some aspects of HOS estimators [2, 15] are not followed up here.
KeywordsTail Length Arithmetic Mean Truncation Parameter Little Trim Square Generalise Gaussian Distribution
Unable to display preview. Download preview PDF.
- W. Göhler. Höhere Mathematik. Dt. Verl. für Grundstoffindustrie, Leipzig, 1987.Google Scholar
- F. R. Hampel. The influence curve and its role in robust estimation, volume 69, pages 383–393. J. Am. Stat. Assoc., 1974.Google Scholar
- G. C. W. Leung and D. Hatzinakos. Efficient implementations of higher— order statistics estimators. Proc IEEE Signal Processing Workshop on Higher Order Statistics, pages 266– 270, Jun 1995.Google Scholar
- A. K. Nandi. Robust estimation of third-order cumulants in applications of higher-order statistics. IEE Proceedings Part F, 140(6):380–389, 1993.Google Scholar
- A. K. Nandi and D. Mämpel. On robust estimation of cumulants. In M. J. J. Holt, C. F. N. Cowan, P. M. Grant, and W. A. Sandham, editors, Proceedings of EUSIPCO’ 94, pages 470– 473, 1994.Google Scholar
- A. Stuart and J. K. Ord. Kendall’ s Advanced Theory of Statistics, volume I. Edward Arnold, London, 6th edition, 1994.Google Scholar
- Y. Zhang, D. Hatzinakos, and A. N. Venetsanopoulos. Bootstrapping techniques in the estimation of higher-order cumulants from short data records. In Proceedings of ICASSP’ 93, volume 4, pages 200–203, Minneapolis, 1993.Google Scholar