Revistes Publicacions URV: SORT - Statistics and Operations Research Transactions> 2016

Log-ratio methods in mixture models for compositional data sets

  • Identification data

    Identifier: RP:2450
    Handle: https://hdl.handle.net/20.500.11797/RP2450
    Authors:
    Mateu-Figueras, GlòriaMartín-Fernández, Josep AntoniComas-Cufí, Marc
    Abstract:
    When traditional methods are applied to compositional data misleading and incoherent results could be obtained. Finite mixtures of multivariate distributions are becoming increasingly important nowadays. In this paper, traditional strategies to fit a mixture model into compositional data sets are revisited and the major difficulties are detailed. A new proposal using a mixture of distributions defined on orthonormal log-ratio coordinates is introduced. A real data set analysis is presented to illustrate and compare the different methodologies.

  • Others:

    URV's Author/s: Mateu-Figueras, Glòria Martín-Fernández, Josep Antoni Comas-Cufí, Marc
    Keywords: Compositional data, Finite Mixture, Log ratio, Model-based clustering, Normal distribution, Orthonormal coordinates, Simplex
    Abstract: When traditional methods are applied to compositional data misleading and incoherent results could be obtained. Finite mixtures of multivariate distributions are becoming increasingly important nowadays. In this paper, traditional strategies to fit a mixture model into compositional data sets are revisited and the major difficulties are detailed. A new proposal using a mixture of distributions defined on orthonormal log-ratio coordinates is introduced. A real data set analysis is presented to illustrate and compare the different methodologies.
    Journal publication year: 2016
    Publication Type: info:eu-repo/semantics/publishedVersion info:eu-repo/semantics/article

  • Keywords:

    Compositional data, Finite Mixture, Log ratio, Model-based clustering, Normal distribution, Orthonormal coordinates, Simplex

  • Documents:

    DocumentPrincipal

  • Cerca a google

    Search to google scholar