基于MCEM算法的多元正态分布均值向量估计

Mean vector estimation of multivariate normal distribution based on the MCEM algorithm

将Monte Carlo EM算法应用到存在缺失数据的多元正态分布数据中,给出了其均值向量的估计公式,并利用R语言进行了不同样本量下的随机模拟,最后将模拟结果与传统均值插补方法进行了对比,结果表明MCEM算法估计效果相比于传统均值插补方法更好.

The Monte Carlo EM algorithm is used in multivariate normal distribution data with missing data,and the estimation formula of its mean vector is given.The random simulation under different sample size is carried out by using R programming language.Finally,the simulation results are compared with the traditional mean imputation method,and the results show that the MCEM algorithm is more accurate than the traditional method.