Robust and Efficient One-way MANOVA Tests with Fast, Robust Bootstrap
We provide MATLAB code to perform robust and efficient One-way MANOVA Tests based on multivariate S or MM-estimators of the centers and common shape of the groups. The inference is
based on the fast and robust bootstrap. Questions regarding the code can be sent to
Gert Willems.
MATLAB code
Function to compute robust MANOVA test based on S-estimates of the centers and shape with
inference based on the fast and robust bootstrap: robWLS_FRB.m
Function to compute robust MANOVA test based on MM-estimates of the centers and shape with
inference based on the fast and robust bootstrap: robWLMM_FRB.m
Necessary additional function that computes the one-sample multivariate S-estimates of
location and shape: multiS.m
Necessary additional function that computes the k-sample multivariate S-estimates of
location and shape: multiSg.m
Necessary additional function that computes the one-sample multivariate MM-estimates of
location and shape: multiMM_c.m
Necessary additional function that computes the k-sample multivariate MM-estimates of
location and shape: multiMMg_c.m
Necessary additional functions that execute the fast and robust
bootstrap for S-estimates:
RobBoot_S.m
RobBoot_Sg.m
Necessary additional functions that execute the fast and robust
bootstrap for MM-estimates:
RobBoot_MM_c1.m
RobBoot_MMg_c1.m
Reference
Van Aelst, S., and Willems, G. (2011), "
Robust and Efficient One-way MANOVA Tests, "
Journal of the American Statistical Association , 106, 706-718.Technical report