Murphy, Thomas Brendan

Thumbnail Image
Preferred name
Murphy, Thomas Brendan
Official Name
Murphy, Thomas Brendan

Research Output

Filters

2
2
Reset filters

Settings

Now showing 1 - 2 of 2
  • Publication
    Variable selection methods for model-based clustering
    (The American Statistical Association, the Bernoulli Society, the Institute of Mathematical Statistics, and the Statistical Society of Canada, 2018-04-26)
    Fop, Michael 
    ;
    Murphy, Thomas Brendan 
    Model-based clustering is a popular approach for clustering multivariate data which has seen applications in numerous fields. Nowadays, high-dimensional data are more and more common and the model-based clustering approach has adapted to deal with the increasing dimensionality. In particular, the development of variable selection techniques has received a lot of attention and research effort in recent years. Even for small size problems, variable selection has been advocated to facilitate the interpretation of the clustering results. This review provides a summary of the methods developed for variable selection in model-based clustering. Existing R packages implementing the different methods are indicated and illustrated in application to two data analysis examples.
      496Scopus© Citations 60
  • Publication
    Variable Selection for Latent Class Analysis with Application to Low Back Pain Diagnosis
    (The Institute of Mathematical Statistics, 2017-12-28)
    Fop, Michael 
    ;
    Smart, Keith 
    ;
    Murphy, Thomas Brendan 
    The identification of most relevant clinical criteria related to low back pain disordersis a crucial task for a quick and correct diagnosis of the nature of pain and its treatment.Data concerning low back pain can be of categorical nature, in form of check-list in whicheach item denotes presence or absence of a clinical condition. Latent class analysis is amodel-based clustering method for multivariate categorical responses which can be appliedto such data for a preliminary diagnosis of the type of pain. In this work we propose avariable selection method for latent class analysis applied to the selection of the mostuseful variables in detecting the group structure in the data. The method is based onthe comparison of two different models and allows the discarding of those variables withno group information and those variables carrying the same information as the alreadyselected ones. We consider a swap-stepwise algorithm where at each step the models arecompared through and approximation to their Bayes factor. The method is applied tothe selection of the clinical criteria most useful for the clustering of patients in differentclasses of pain. It is shown to perform a parsimonious variable selection and to give agood clustering performance. The quality of the approach is also assessed on simulateddata
    Scopus© Citations 33  544