No label defined (Q3976437)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 19011
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | No label defined |
scientific article; zbMATH DE number 19011 |
Statements
26 June 1992
0 references
p-subset algorithm
0 references
multiple regression
0 references
outliers
0 references
robust estimators
0 references
outlier diagnostics
0 references
masking effect
0 references
high breakdown point
0 references
least trimmed squares estimator
0 references
approximate algorithm
0 references
equivariance
0 references
one-dimensional location model
0 references
half sample
0 references
0.8486812114715576
0 references
0.8360693454742432
0 references
0.8303157687187195
0 references
0.8289920091629028
0 references
[For the entire collection see Zbl 0731.00019.] In multiple regression outliers cannot be identified by most of robust estimators and outlier diagnostics if several of them form a cluster. To prevent such a masking effect one needs a method that can cope with a large fraction of outliers. One such method with high breakdown point is the least trimmed squares estimator. Its exact computation, however, is quite difficult because the objective function may have a large number of local minima. The authors use an approximate algorithm based on p-subsets and prove that the algorithm shares the equivariance and good breakdown properties of the exact estimator. The same result is valid for other high- breakdown-point estimators. Finally, for the one-dimensional location model special results are given concerning half samples.
0 references
