ALTERNATIVE ROBUST ESTIMATORS FOR PARAMETERS OF THE LINEAR REGRESSION MODEL

This paper considers parameter estimation of the linear regression model with Ramsay-Novick (RN) distributed errors, focusing on its use as an aid to robustness. Positioning within class heavy tailed distributions, RN distribution can be defined modification unbounded influence function a non-robust density so that it has more resistance outliers. Potential this robust have far been assessed in Bayesian settings real data examples and there is lack performance assessment for finite samples classical approach. study therefore explores robustness properties when used error comparison normal well other alternating heavy-tailed distributions like Laplace Student-t. An extensive simulation was conducted purpose under different sample size, parameters outlier percantages. efficient generation through random-walk Metropolis algorithm here also suggested. The results were supported by world application famously known Brownlee’s stack loss plant data.