Edinburgh Research Explorer

Bayesian Machine Learning Approaches for Longitudinal Latent Class Modelling to Define Wheezing Phenotypes to Elucidate Environmental Associates

Research output: Contribution to journalArticle

Related Edinburgh Organisations

Original languageEnglish
Number of pages4
JournalQuaderni di statistica
Volume14
Issue number14
Publication statusPublished - 2012

Abstract

Accurate phenotypic definition of wheezing in childhood can lead to a greater understanding of the distinct physiological markers associated with different wheeze phenotypes. This paper looks at Bayesian machine learning approaches using Infer.NET to define wheeze phenotypes based on both parental questionnaires and General Practitioner data on patterns of asthma and wheeze consultation within the first eight years of life. We illustrate a taxonomy of longitudinal latent class item response models with varying modelling assumptions to determine wheeze phenotypes (latent classes) for homogenous groups of children.

ID: 22927596