作者: Ana M. Bianco , Graciela Boente , Isabel M. Rodrigues
DOI: 10.1080/03610926.2015.1026997
关键词:
摘要: ABSTRACTThis paper presents a procedure for testing the hypothesis that underlying distribution of data is elliptical when using robust location and scatter estimators instead sample mean covariance matrix. Under mild assumptions include distributions without first moments, we derive test statistic asymptotic behavior under null special alternatives. Numerical experiments allow to compare tests based on matrix with estimators, various different We also provide numerical comparison other competing tests.