Least Trimmed Squares: Cointegration and Outliers
研究了在协整自回归分布滞后模型中，使用最小修剪平方估计量系统处理异常值的方法，证明其渐近性质与基于好误差的普通最小二乘估计相同。
ABSTRACT When applying the cointegrated autoregressive distributed lag model it is common to include indicator variables for outliers. This is often done in a somewhat ad hoc way. Least Trimmed Squares estimation provides a more systematic approach. This estimator is robust to a large number of outliers of many types. We analyse the estimator in a model that allows a range of contamination and show that it has the same asymptotic properties as the infeasible Ordinary Least Squares estimator applied to a model generated by the good errors.