Weighted nonparametric regression estimation with truncated and dependent data

By applying the empirical likelihood method, we construct a new weighted Nadaraya-Watson type estimator of the conditional mean function for a left truncation model. The function includes the regression function, conditional moment as well as conditional distribution function. Under strong mixing assumptions, we obtain the asymptotic normality and weak consistency of the estimator. Finite sample behaviour of the estimator is investigated via simulations too.