You are here

Examining the effect of initialization strategies on the performance of Gaussian mixture modeling.

Title: Examining the effect of initialization strategies on the performance of Gaussian mixture modeling.
3 views
0 downloads
Name(s): Shireman, Emilie, author
Steinley, Douglas, author
Brusco, Michael J, author
Type of Resource: text
Genre: Journal Article
Text
Date Issued: 2017-02-01
Physical Form: computer
online resource
Extent: 1 online resource
Language(s): English
Abstract/Description: Mixture modeling is a popular technique for identifying unobserved subpopulations (e.g., components) within a data set, with Gaussian (normal) mixture modeling being the form most widely used. Generally, the parameters of these Gaussian mixtures cannot be estimated in closed form, so estimates are typically obtained via an iterative process. The most common estimation procedure is maximum likelihood via the expectation-maximization (EM) algorithm. Like many approaches for identifying subpopulations, finite mixture modeling can suffer from locally optimal solutions, and the final parameter estimates are dependent on the initial starting values of the EM algorithm. Initial values have been shown to significantly impact the quality of the solution, and researchers have proposed several approaches for selecting the set of starting values. Five techniques for obtaining starting values that are implemented in popular software packages are compared. Their performances are assessed in terms of the following four measures: (1) the ability to find the best observed solution, (2) settling on a solution that classifies observations correctly, (3) the number of local solutions found by each technique, and (4) the speed at which the start values are obtained. On the basis of these results, a set of recommendations is provided to the user.
Identifier: FSU_pmch_26721666 (IID), 10.3758/s13428-015-0697-6 (DOI), PMC4930421 (PMCID), 26721666 (RID), 26721666 (EID), 10.3758/s13428-015-0697-6 (PII)
Keywords: Algorithm initialization, EM algorithm, Mixture modeling
Grant Number: R01 AA023248
Publication Note: This NIH-funded author manuscript originally appeared in PubMed Central at https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4930421.
Subject(s): Algorithms
Finite Element Analysis
Models, Theoretical
Normal Distribution
Probability
Persistent Link to This Record: http://purl.flvc.org/fsu/fd/FSU_pmch_26721666
Owner Institution: FSU
Is Part Of: Behavior research methods.
1554-3528
Issue: iss. 1, vol. 49

Choose the citation style.
Shireman, E., Steinley, D., & Brusco, M. J. (2017). Examining the effect of initialization strategies on the performance of Gaussian mixture modeling. Behavior Research Methods. Retrieved from http://purl.flvc.org/fsu/fd/FSU_pmch_26721666