초록 close

Consider the regression modelY_i=g(x_i)+e_i for i=1,2,ldots, n,where: (1) x_i are fixed design points, (2) e_i are independentrandom errors with mean zero, (3) g(cdot) is unknown regressionfunction defined on [0,1]. Under Y_i are censored randomly, wediscuss the asymptotic normality of the weighted kernel estimatorsof g when the censored distribution function is known or unknown.