Rank regression methods for left-truncated and right-censored data (Q1175376)

This paper considers the problem of estimation of regression coefficients when the response variable is right truncated and left censored. Rank estimators and adaptive rank estimators are suggested and their consistency and asymptotic normality is proved under suitable regularity conditions using martingale theory and a tightness lemma for stochastic integrals of empirical processes. An extension to the multiple regression case is also presented.