The interval testing procedure: A general framework for inference in functional data analysis.
Biometrics
; 72(3): 835-45, 2016 09.
Article
en En
| MEDLINE
| ID: mdl-26811864
ABSTRACT
We introduce in this work the Interval Testing Procedure (ITP), a novel inferential technique for functional data. The procedure can be used to test different functional hypotheses, e.g., distributional equality between two or more functional populations, equality of mean function of a functional population to a reference. ITP involves three steps:
(i) the representation of data on a (possibly high-dimensional) functional basis; (ii) the test of each possible set of consecutive basis coefficients; (iii) the computation of the adjusted p-values associated to each basis component, by means of a new strategy here proposed. We define a new type of error control, the interval-wise control of the family wise error rate, particularly suited for functional data. We show that ITP is provided with such a control. A simulation study comparing ITP with other testing procedures is reported. ITP is then applied to the analysis of hemodynamical features involved with cerebral aneurysm pathology. ITP is implemented in the fdatest R package.Palabras clave
Texto completo:
1
Banco de datos:
MEDLINE
Asunto principal:
Interpretación Estadística de Datos
/
Modelos Estadísticos
Tipo de estudio:
Risk_factors_studies
Límite:
Humans
Idioma:
En
Revista:
Biometrics
Año:
2016
Tipo del documento:
Article
País de afiliación:
Italia