Department of Mathematical Sciences
|DATE:||Thursday, March 13, 2020|
|TIME:||1:15pm – 2:15pm|
|SPEAKER:||Mengyu Chen, Binghamton University|
|TITLE:||Maximum Empirical Likelihood Estimators for Linear Regression Models with Symmetric Errors|
When dealing with linear regression models, there are two common approaches to estimate the parameters. One is the least squares estimator and the other one is the least absolute deviation estimators.
For each estimator, there is a corresponding criterion function. LS estimator is the MLE of parameters when the error term is normally distributed, and LAD estimator is the MLE when the error term follows Laplace distribution. For other distributions, those criterion functions may not be enough. To make more precise estimation, we can combine those two criterion functions. The comparison of these three criterion functions will be discussed.