Earth and Planetary Sciences
9 years ago

Mixed estimator outperforms least square estimator in regression coefficient estimation.

2 views
JUSTC
Small Sample Properties for Mixed Estimator of Regression Coefficients Under Misspecified Regression Model
Wei Lai-sheng

Paper Summary

The mixed regression estimator (MRE) outperforms the least square estimator (LSE) in estimating regression coefficients when the regression model is not perfectly specified. This was shown through comparisons based on the mean square error matrix (MSE M) and Pitman Closeness (PC) criteria.