MatTransMix: an R Package for Matrix Model-Based Clustering and Parsimonious Mixture Modeling

Zhu, Xuwen; Sarkar, Shuchismita; Melnykov, Volodymyr

doi:10.1007/s00357-021-09401-9

Cited by 10 publications

(3 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The cluster analysis have been performed with the package (Zhu et al, 2022 ) of the statistical software R. As usual when performing clustering, the main parameter to set is represented by the number of clusters K . Moreover, it is important that the clusters are interpretable (Fraley & Raftery, 1998 ; Forgy, 1965 ).…”

Section: Analysis and Resultsmentioning

confidence: 99%

“…Matrix-variate models suffer from over-parametrization that leads to estimation issues. This issue is addressed in Sarkar et al ( 2020 ) and Zhu et al ( 2022 ), with the aim to explain the data with as few parameters as possible. To do so, the spectral decomposition of the covariance matrix (Banfield & Raftery, 1993 ; Celeux & Govaert, 1995 ) is used.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

A Comparison of Migrant Integration Policies via Mixture of Matrix-Normals

et al. 2022

View full text Add to dashboard Cite

In recent decades, there has been a growing interest in comparative studies about migrant integration, assimilation and the evaluation of policies implemented for these purposes. Over the years, the Migrant Integration Policy Index (MIPEX) has become a reference on these topics. This index measures and evaluates the policies of migrants’ integration in 52 countries over time. However, the comparison of very different countries can be difficult and, if not well conducted, can lead to misleading interpretations and evaluations of the results. The aim of this paper is to improve this comparison and facilitate the reading of the considered phenomenon, by applying a Mixture of Matrix-Normals classification model for longitudinal data. Focusing on data for 7 MIPEX dimensions from 2014 to 2019, our analysis identify 5 clusters of countries, facilitating the evaluation and the comparison of the countries within each cluster and between different clusters.

show abstract

Section: Analysis and Resultsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

A Comparison of Migrant Integration Policies via Mixture of Matrix-Normals

et al. 2022

View full text Add to dashboard Cite

show abstract

“…A complete analysis of the EM algorithm, for all the possible parsimonious combinations of model , can be found in Sarkar et al. (2020) and Zhu & Melnykov (2020). From a computational point of view, we use a random start strategy.…”

Section: Methodsmentioning

confidence: 99%

Modelling students’ career indicators via mixtures of parsimonious matrix‐normal distributions

Tomarchio

Ingrassia

Melnykov

2022

Aus NZ J of Statistics

View full text Add to dashboard Cite

Summary The evaluation of the teaching efficiency, under different points of view, is an important aspect for the university system because it helps managers to improve more and more the quality of the education and helps students to achieve strong professional skills. In this framework, students’ careers as well as teachers’ qualification and quantity adequacy indicators are analysed based on data sets provided by the Italian National Agency for the Evaluation of Universities and Research Institutes (ANVUR) according to a mixture model approach. In particular, parsimonious mixtures of matrix‐normal distributions are used to detect underlying grouping structures. The results show that the data present an underlying group structure of courses having different traits, thus providing useful information for the university policy makers.

show abstract

Parsimonious Finite Mixtures of Matrix-Variate Regressions

Punzo

Tomarchio

2022

Emerging Topics in Statistics and Biostatistics

View full text Add to dashboard Cite

MatTransMix: an R Package for Matrix Model-Based Clustering and Parsimonious Mixture Modeling

Cited by 10 publications

References 21 publications

A Comparison of Migrant Integration Policies via Mixture of Matrix-Normals

A Comparison of Migrant Integration Policies via Mixture of Matrix-Normals

Modelling students’ career indicators via mixtures of parsimonious matrix‐normal distributions

Parsimonious Finite Mixtures of Matrix-Variate Regressions

Contact Info

Product

Resources

About