Random Start Forward Searches for Detecting Mixtures of Regression Models
To detect outliers from a single regression model requires one, perhaps robust, fit to the data. But if the “outlying observations” are other regression models, it may be necessary to fit several different linear models in order to reveal the structure. We illustrate the diagnostic
use of random start forward searches to reveal mixtures of regression models.
ATKINSON Anthony C.;
RIANI Marco;
CERIOLI Andrea;
PERROTTA Domenico;
2014-10-20
The International Statistical Institute/International Association for Statistical Computing
JRC89156
978-2-8399-1347-8,
https://publications.jrc.ec.europa.eu/repository/handle/JRC89156,
Additional supporting files
| File name | Description | File type | |
|---|
