Initializing the EM algorithm in Gaussian mixture models with an unknown number of components
Initializing the EM algorithm in Gaussian mixture models with an unknown number of components
复制标题
DOI：
10.1016/j.csda.2011.11.002
复制
发表时间：
2012-06-01
影响因子：
1.8
通讯作者：
Melnykov, Igor
中科院分区：
文献类型：
--
作者：
Melnykov, Volodymyr;Melnykov, Igor
提出了一种在分量数目未知的多元高斯混合模型中初始化期望最大化(EM)算法的方法。由于EM算法往往对初始参数向量的选择非常敏感，有效的初始化是算法未来收敛到似然函数的最优局部极大值的一个重要的预备过程。我们提出了一种通过选择邻域密度较高的点来初始化均值向量的策略，并使用截断正态分布来初步估计离散度矩阵。通过实例说明了该方法，并与其他几种初始化方法进行了比较。(C)2011爱思唯尔B.V.保留所有权利。
An approach is proposed for initializing the expectation-maximization (EM) algorithm in multivariate Gaussian mixture models with an unknown number of components. As the EM algorithm is often sensitive to the choice of the initial parameter vector, efficient initialization is an important preliminary process for the future convergence of the algorithm to the best local maximum of the likelihood function. We propose a strategy initializing mean vectors by choosing points with higher concentrations of neighbors and using a truncated normal distribution for the preliminary estimation of dispersion matrices. The suggested approach is illustrated on examples and compared with several other initialization methods. (C) 2011 Elsevier B.V. All rights reserved.