Robust two-group discrimination by bounded influence regression: a Monte Carlo simulation
Computational Statistics & Data Analysis
High breakdown estimation for multiple populations with applications to discriminant analysis
Journal of Multivariate Analysis
Computational Statistics & Data Analysis
A new statistic in the one-way multivariate analysis of variance
Computational Statistics & Data Analysis
Editorial: Special issue on variable selection and robust procedures
Computational Statistics & Data Analysis
Error rates for multivariate outlier detection
Computational Statistics & Data Analysis
Hi-index | 0.03 |
The Wilks' Lambda Statistic (likelihood ratio test, LRT) is a commonly used tool for inference about the mean vectors of several multivariate normal populations. However, it is well known that the Wilks' Lambda statistic which is based on the classical normal theory estimates of generalized dispersions, is extremely sensitive to the influence of outliers. A robust multivariate statistic for the one-way MANOVA based on the Minimum Covariance Determinant (MCD) estimator will be presented. The classical Wilks' Lambda statistic is modified into a robust one through substituting the classical estimates by the highly robust and efficient reweighted MCD estimates. Monte Carlo simulations are used to evaluate the performance of the test statistic under various distributions in terms of the simulated significance levels, its power functions and robustness. The power of the robust and classical statistics is compared using size-power curves, for the construction of which no knowledge about the distribution of the statistics is necessary. As a real data application the mean vectors of an ecogeochemical data set are examined.