截尾线性回归模型参数极大似然估计的EM算法

对区间截尾的线性回归模型,在误差服从正态分布的条件下,给出其基于似然方程的参数极大似然估计(Maximum Likelihood Estimate,MLE)的一般迭代算法,证明了EM算法与该一般迭代算法的一致性,由EM算法的性质保证了迭代的收敛性,证明了Murray Aitkin给出的右截尾数据回归模型参数MLE的EM算法是本文的一个特例以及EM算法收敛点的唯一性.

EM Algorithm for the Maximum Likelihood Estimate of the Parameters in Censored Linear Regression Models

For linear regression models with interval censored,the authors have presented an ordinary iterative algorithm for the maximum likelihood estimate(MLE) on the parameters of likelihood-based equations when errors are normally distributed,and have proved the consistency between the iterative algorithm and EM algorithm,the nature of EM algorithm ensuring iterative convergence.The authors have pointed out that,the EM algorithm for the MLE on the parameters in right-censored linear regression models proposed by Murray Aitkin is merely a special case of the EM algorithm as proposed in this paper.In the end,the authors have proved the uniqueness of convergence point of EM algorithm.