The paper contains some results on model selection in conditional independence graphs when the variables have an extended Skew Normal distribution. This family is a slight modification of the Skew Normal one, that extends the class of Normal distributions through the addition of a shape parameter that regulates the skewness. A test based on a Wald-type statistic for a single edge exclusion/inclusion is proposed. Its performances, in finite samples, are assessed through numerical experiments.
Capitanio A., Pacillo S. (2008). A Wald's test for conditional independence Skew Normal graphs. HEIDELBERG : Physica-Verlag.
A Wald's test for conditional independence Skew Normal graphs
CAPITANIO, ANTONELLA;
2008
Abstract
The paper contains some results on model selection in conditional independence graphs when the variables have an extended Skew Normal distribution. This family is a slight modification of the Skew Normal one, that extends the class of Normal distributions through the addition of a shape parameter that regulates the skewness. A test based on a Wald-type statistic for a single edge exclusion/inclusion is proposed. Its performances, in finite samples, are assessed through numerical experiments.File in questo prodotto:
Eventuali allegati, non sono esposti
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
