Maximum Likelihood Estimation Under Linear Inequality Restrictions In Cox Proportional Hazards Model With Missing Covariates

This paper discusses maximum likelihood estimation in Cox Proportional Hazards Model under linear inequality restrictions with missing covariates. In practice, censored data, truncted data and misisng data are often encountered with. It is useful to use the incomplete data information and restriction information of parameters. This paper uses EM algorithm and constrained optimation algorithm to obntain the maximum likelihood estimates under linear inequality restrictions with missing data.