NotesFAQContact Us
Search Tips
ERIC Number: ED406393
Record Type: Non-Journal
Publication Date: 1996
Pages: 16
Abstractor: N/A
Reference Count: N/A
High-Breakdown Regression by Least Quartile Differences.
Blankmeyer, Eric
A high-breakdown estimator is a robust statistic that can withstand a large amount of contaminated data. In linear regression, high-breakdown estimators can detect outliers and distinguish between good and bad leverage points. This paper summarizes the case for high-breakdown regression and emphasizes the least quartile difference estimator (LQD) proposed by C. Croux, P. J. Rousseeuw, and O. Hossjer (1994). This regression method examines the absolute differences between every pair of residuals and minimizes the first quartile of these differences with an adjustment for degrees of freedom. LQD is affine equivalent and has a 50% breakdown point, the highest possible. Its asymptotic efficiency is about 67% of least-squares, so LQD should be able to deal with anomalous observations and should also perform well when the data are not contaminated. Although interest in the approach is growing, software is still not widely available. An appendix presents a BASIC computer program for one type of high-breakdown regression. (Contains 17 references.) (SLD)
Publication Type: Reports - Evaluative; Computer Programs
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A