ABSTRACT
We now introduce covariates into a time to event analysis via the accelerated failure time (AFT) model and the proportional hazards (PH) model. As in Chapter 12, we have a sample of data of the form D = (y,δ ) with y = (y1, . . . ,yn)′, where yi = min(Ti,Ci), and δ = (δ1, . . . ,δn)′, where δi = I[0,∞)(Ci − Ti) ≡ I(Ti ≤ Ci). The censoring times are assumed to be noninformative and to be independent of event times. In addition, we now have xi, an r vector of predictor information associated with individual i.
