区间删失数据下Weibull 比例优势模型的参数估计

Parameter estimation of Weibull proportional odds model under interval-censored data

基于区间删失数据建模是当前复杂数据分析的热点之一。本研究在两类区间删失数据下建立Weibull比例优势模型,基于极大似然估计给出了模型参数,进一步讨论了估计量的渐近性质。数值模拟验证了模型参数的估计效果,并将提出的模型及方法应用到艾滋病临床试验数据和肺肿瘤试验数据中,给出了感兴趣事件的生存函数曲线,通过与生存函数的非参数极大似然估计比较,表明所提方法具有较好的拟合效果。

Modeling based on interval-censored data is one of the hotspots of complex data analysis at present.In this paper,the Weibull proportional odds model was established under the two types of interval-censored data.The parameter estimation of the model was given based on the maximum likelihood estimation and the asymptotic normality of the estimator was further discussed.Numerical simulation verified the estimation effect of model parameters.Then the proposed model was applied to AIDS clinical trial data and lung tumor experimental data,and the smooth survival function curve was drawn.Compared with the nonparametric maximum likelihood estimation of survival function,the proposed method has been proven to have better fitting effect.