NotesFAQContact Us
Search Tips
ERIC Number: ED411271
Record Type: Non-Journal
Publication Date: 1997-Mar
Pages: 38
Abstractor: N/A
Reference Count: N/A
A Weighted Least Squares Approach To Robustify Least Squares Estimates.
Lin, Chowhong; Davenport, Ernest C., Jr.
This study developed a robust linear regression technique based on the idea of weighted least squares. In this technique, a subsample of the full data of interest is drawn, based on a measure of distance, and an initial set of regression coefficients is calculated. The rest of the data points are then taken into the subsample, one after another, and a weighted least squares procedure is performed each time a new data point is brought in, until all data points are included. The weighted average and standard errors of the regression coefficients from all iterations are calculated and compared with those from ordinary least squares and two other robust techniques. It is shown that the technique developed in this paper performs better than the least absolute deviation approach and at the same level with the least median of squares (LMS) approach. The simplicity of this approach, however, seems to justify its use over the LMS. (Contains 5 figures, 10 tables, and 8 references.) (Author/SLD)
Publication Type: Reports - Evaluative; Speeches/Meeting Papers
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A