Partially Linear Additive Hazards Regression for Bivariate Interval-Censored Data

In this paper, we discuss regression analysis of bivariate interval-censored failure time data that often occur in biomedical and epidemiological studies. To solve problem, propose a kind general flexible copula-based semiparametric partly linear additive hazards models can allow for both time-dependent covariates possible nonlinear effects. For inference, sieve maximum likelihood estimation approach based on Bernstein polynomials is proposed to estimate the baseline hazard functions covariate The resulting estimators parameters are shown be consistent, asymptotically efficient normal. A simulation study conducted assess finite-sample performance method results show it effective practice. Moreover, an illustration provided.