scientific article; zbMATH DE number 4068074
[For the entire collection see Zbl 0627.00024.] There exist two different possibilities for the modelling of outlier situations in statistical data: In the first case n-k observations from p.d. F are contaminated by k observations coming from p.d. \(G\neq F\). In the second case a p.d. produces outlying observations from itself owing to a heavy tailed distribution function. Some of the models of the first kind are presented. Then it is shown that the definition of the outlier behaviour of single distributions by \textit{R. F. Green} [J. Am. Stat. Assoc. 71, 502-505 (1976; Zbl 0344.62016)] can be reformulated in a non- asymptotic manner with respect to the sample size n.
This page was built for publication:
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3800902)