In this article, an estimation procedure to simple linear regression in the presence of outliers is proposed. The performance of the proposed estimator, the AM estimator, is compared with other traditional estimators: least squares, Theil type repeated median, and geometric mean. A numerical example is given to illustrate the proposed estimator. Simulation results indicate that the proposed estimator is accurate and has a high precision in the presence of outliers.
Ebrahem, Mohammed Al-Haj and Al-Nasser, Amjad D.
"Estimating The Slope Of Simple Linear Regression In The Presence Of Outliers,"
Journal of Modern Applied Statistical Methods:
2, Article 15.
Available at: http://digitalcommons.wayne.edu/jmasm/vol4/iss2/15