| Summary: | 碩士 === 靜宜大學 === 財務與計算數學系 === 100 === The multivariate current status failure time data consist several possibly related event times of interest, in which the status of each event is determined at an examination time. If the examination time is intrinsically related to the event times, the examination is referred to as informative censoring and needed to be taken into account. Such data often occur in, for example, epidemiological survey, cancer research and animal carcinogenicity experiment. This thesis proposes a frailty model, which characterizes the correlation among the event times by a shared random effect. The frailty also accounts for the informative censoring simultaneously. Likelihood approach is proposed, in which the likelihood is approximated by the Gaussian quadrature techniques. Thus, maximum likelihood estimation is derived. To investigate finite sample properties of the proposed method, extensive simulation studies are conducted.
|
|---|
