Weighted Least Trimmed Estimator


Version: 1.2.0


Download

DESCRIPTION:
The package computes the WLTE(k) for the given set of observations. One of the advantages of these estimators is that they posses a high breakdown point properties. This allows an estimate, robust to the presence of outliers, to be obtained.

The package is written and work under MATLAB.

AUTHORS:
Dimitar Atanasov
datanasov@fmi.uni-sofia.bg

LICENSE:
The package is under the CC-GNU LGPL.
You may use, copy or redistribute this package free if name of the package, names and e-mail addresses of the authors are cited.

DOWNLOAD:

REFERENCES:

  1. Atanasov D. (2003a) About the Concept of Weights of WLTE(k) Estimators. Pliska Studia Matematica Bulgaria. Vol. 14. 5-14.
  2. Atanasov D.V., Neykov N.M. (2001). On the Finite Sample Breakdown Point of the WLTE(k) and d-fullness of a Set of Continuous Functions, In: Proceedings of the VI International Conference ''Computer Data Analysis And Modeling'', Minsk, Belarus. 52-57.
  3. Neykov N.M., Neytchev, P.N. (1990). A Robust Alternative of the Maximum Likelihood Estimators, COMPSTAT 1990, Short Communications, 99 - 100.
  4. Neykov N., Muller Ch. (2003). Breakdown Points and computation of trimmed likelihood estimators in generalized linear models. Development in Robust Statistics, R. Dutter et al, (eds.). Physica - Verlag. Heidelberg. 503 - 519.
  5. Vandev D.L. (1993). A Note on Breakdown Point of the Least Median of Squares and Least Trimmed Estimators. Statistics andProbability Letters. Vol. 16, 117 - 119.
  6. Vandev D.L. (2005). Stochastic Optimization in Robust Statistics. Pliska Studia Matematica Bulgaria. Vol. 17. 323 - 337.
  7. Vandev D.L. and Neykov, N.M. (1998). About Regression Estimators with High Breakdown Point, Statistics. 111 - 129.