The modified maximum likelihood regression type estimators using bivariate ranked set sampling

Rank set sampling (RSS) is an efficient sampling technique, used especially in the situations where the measurement on the variable of interest is time-consuming, costly or difficult. One of the modifications to make it more efficient is the regression type estimation using bivariate RSS. We derive the modified maximum likelihood (MML) regression type estimators using bivariate RSS when the concomitant variable X is stochastic and the error term is non-normal where it is problematic to obtain the maximum likelihood (ML) estimators. The procedures and merits of the proposed estimators are illustrated through simulations and two real-life applications.