Abstract
An approach is proposed for initializing the expectationmaximization (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.
Original language | English (US) |
---|---|
Pages (from-to) | 1381-1395 |
Number of pages | 15 |
Journal | Computational Statistics and Data Analysis |
Volume | 56 |
Issue number | 6 |
DOIs | |
State | Published - Jun 2012 |
Externally published | Yes |
Bibliographical note
Copyright:Copyright 2012 Elsevier B.V., All rights reserved.
Keywords
- EM algorithm
- Eigenvalue decomposition
- Gaussian mixture model
- Initialization
- Truncated normal distribution