科技广场Issue(2)：6-10,5.
EM算法在混合正态分布模型参数估计中的应用研究
Application Research of EM Algorithm in Parameter Estimation of the Mixture Normal Distribution Model
摘要
Abstract
Finite mixtures of distributions have provided an approach to the statistical modeling of a wide va-riety of random phenomena. Many random phenomena can be described by normal distribution when their sample sizes are large enough. Mixture normal distribution is widely used in this situation for its flexible and efficient fit-ting ability. In this paper, the parameters of mixture normal distribution are estimated by EM algorithm. The opti-mizations of complex likelihood functions are turned into a series of optimizations of functions by data expansion in this algorithm, which can obtain the corresponding iterative formula. Finally, the simulation results illustrated the algorithm's convergence and effectiveness.关键词
混合正态分布模型/EM算法/极大似然估计Key words
Mixture Normal Distribution Model/EM Algorithm/Maximum Likelihood Estimation分类
数理科学引用本文复制引用
杨晴,魏立力..EM算法在混合正态分布模型参数估计中的应用研究[J].科技广场,2014,(2):6-10,5.基金项目
国家自然科学基金项目(编号11261044) （编号11261044）
