Weighted censored quantile regression
Section 2. Estimation of weights using empirical likelihood
We develop a method that converts the auxiliary information to the EL based data driven probabilities, which are further used in the weighted censored quantile regression as the weights. Qin and Lawless (1994) developed the EL approach based on a set of estimating equations. Let be the observed data and the available auxiliary information is represented by the estimating function with parameter, which is known. Then, the maximum empirical likelihood is given by
where and is the parameter in the auxiliary information which can be assumed to be known. The parameter, could be any parametric information available from the previous studies which has an influence on the model parameter, For a given should satisfy to avoid the non-existence of solutions due to convex hull issues. This is the scenario for when zero is not an inner point of the convex hull of the which will fail to provide positive For a given value of using the Lagrange multiplier method, attains its maximum at
The Lagrange multiplier, is the solution to the equation
The estimated are used as the weights in (1.2) for the weighted censored quantile regression. In some cases, may not be available and in such situations, we can use an estimate of say obtained from previous studies. Chen and Qin (1993) showed that for a random sample, and are estimated using (2.2), has smaller variance than the empirical distribution function, As a result, with Bahadur representation (Bahadur, 1966), for a given the quantile estimate, has smaller variance than (See Kuk and Mak, 1989; Rao et al., 1990). Hence our proposed method is expected to be more efficient than the ordinary censored quantile regression.
- Date modified: