Description:
<p>A nonparametric analysis of time-dependent covariate effects on failures determined by a regression function β0(t) in Cox's regression model based on case-cohort sampling design is developed. The analysis is carried out through maximizing appropriate penalized pseudolikelihoods. Weak uniform consistency and pointwise asymptotic normality of the resulting estimators are investigated under regularity conditions. Further generalization of the results is also discussed in the paper.</p>