Inference for the overlap coefficient based on P-splines and Dirichlet process mixtures

• Javier E. Garrido Guillén ^[1] ; Vanda Inácio ; María Xosé Rodríguez Álvarez
  1. [1] University of Edinburgh

     University of Edinburgh

     Reino Unido

     
• Localización: Proceedings of the 35th International Workshop on Statistical Modelling : July 20-24, 2020 Bilbao, Basque Country, Spain / Itziar Irigoyen Garbizu (ed. lit.) , Dae-Jin Lee (ed. lit.) , Joaquín Martínez Minaya (ed. lit.), María Xosé Rodríguez Álvarez (ed. lit.), 2020, ISBN 978-84-1319-267-3, págs. 91-95
• Idioma: inglés
• Enlaces
  • Texto Completo Libro
• Resumen

  • Accurate diagnosis of disease is of great importance in clinical practice and medical research. Before a diagnostic test is routinely used in practice its ability to discriminate between diseased and nondiseased states must be rigorously assessed. Further, its performance may depend on covariates (e.g., age and/or gender). This motivates us to propose the covariate-specific overlap coefficient, which will help to determine the optimal populations where to perform the tests on. We assume a location-scale regression model for the test outcomes in each group, relying on an additive formulation based on Penalised splines, while the regression error follows a Dirichlet process mixture of normal distributions. Our approach is illustrated through an application concerning diagnosis of diabetes.