ABSTRACT
In this chapter we discuss two different empirical likelihood schemes for the Cox proportional hazards regression model. The first approach is to construct the empirical likelihood on the observed data, assuming it comes from a Cox model. The second approach, which we only discuss briefly, is to construct the EL based on the estimating equations given by Cox. Differences and similarities to the Cox partial likelihood are discussed. Joint inference involving the baseline hazard and regression parameters are studied. We also illustrate how the empirical likelihood method can be applied to an extension of the proportional hazards model proposed by Yang and Prentice [129].
