Reconstructive estimation in a parametric random censorship model with incomplete data
The author considers a special case of the model with partially classified data which is a generalization of the well-known competing risks model. The generalization consists of the following: 1) several independent concurrent censoring variables, instead of only one, are considered which are subject to mutual censoring in the sense that any variable can be censored by any other variable; 2) the information about the variable which actually generated the observed data and hence censored the observations of the remaining variables is possibly incomplete. Such a model is referred to in the literature as competing risks model with incomplete information on the cause of failure, or multivariate random censorship model with incomplete data, or model with masked data. The common approach to estimate the unknown parameters of this model consists of deriving the maximum likelihood estimators. Most of the complexity that arises in the solution of the maximum likelihood equation is due to the incompleteness of the information on the components which causes the system failure. This leads to explore possible estimation procedures which allow to reconstruct the missing information and estimate the unknown parameters. The idea of treating the missing data as parameters and maximizing the likelihood over the missing data and parameters is exploited by the author. This approach allows to use the estimates developed for the standard competing risks model thus overcoming the practical impossibility of deriving the estimates in the model with incomplete data. Simulation results are presented. For the entire collection see [Zbl 0990.62506].
- Random partition masking model for censored and masked competing risks data
- Inference based on the EM algorithm for the competing risks model with masked causes of failure
- PARAMETER ESTIMATION IN COX'S REGRESSION MODEL FOR COMPETING RISKS WITH MISSING FAILURE TYPES
- Parameter estimation for partially complete time and type of failure data
This page was built for publication: Reconstructive estimation in a parametric random censorship model with incomplete data
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3843185)