Distributions generated by perturbation of symmetry with emphasis on a multivariate skew t‐distribution
Citations
657 citations
Cites background or methods from "Distributions generated by perturba..."
...The above formulation is in the form presented by Azzalini & Capitanio (2003) and, in a slightly different way, by Genton & Loperfido (2005); despite the discrepancy in publication dates, these two papers have been developed independently and more or less simultaneously....
[...]
...See Fang (2003) and Azzalini & Capitanio (2003) for a detailed analysis of these aspects....
[...]
...More details and some numerical illustrations are given by Azzalini & Capitanio (2003)....
[...]
...The problem has been tackled by Azzalini & Capitanio (2003) where, although a general coincidence could not be established, it has been shown that this is however valid at least for various important cases, notably the multivariate Pearson types II and VII families; the latter family is of special…...
[...]
...The above statements summarize results of Branco & Dey (2001), Gupta (2003) and Azzalini & Capitanio (2003)....
[...]
547 citations
354 citations
Cites background from "Distributions generated by perturba..."
...The multivariate skew t distribution is defined by introducing skewness in a multivariate elliptically symmetric t distribution (26)...
[...]
351 citations
Cites background or methods or result from "Distributions generated by perturba..."
...A third alternative is the skew Student’s t-distribution proposed by Azzalini and Capitanio (2003) (which coincides with the skew t-distribution of Branco and Dey (2001)), having a density on the form...
[...]
...Backtesting shows that the GH skew Student’s t-distribution outperforms the NIG and skew Student’s t-distribution provided by Azzalini and Capitanio (2003) when expected shortfall is used as a risk measure, and is also slightly better at predicting VaR....
[...]
...For Azzalini’s skew Student’s t-distribution, we used the numerical maximum likelihood estimation scheme given in Azzalini and Capitanio (2003) , which has been implemented in the sn-package for R. The CPU time pr. iteration was approximately 0.01 for NIG, approximately 0.02 for GH skew Student’s t and approximately 0.04 for Azzalini’s skew Student’s t, whereas the number of iterations needed until convergence was slightly larger for the GH ......
[...]
...If the data in addition are very skewed, it is also superior to the skew Student’s t-distribution provided by Azzalini and Capitanio (2003) ....
[...]
...For Azzalini’s skew Student’s t-distribution, we used the numerical maximum likelihood estimation scheme given in Azzalini and Capitanio (2003), which has been implemented in the sn-package for R....
[...]
348 citations
Additional excerpts
...In Section 4 we conclude....
[...]
References
2,470 citations
"Distributions generated by perturba..." refers background or methods in this paper
...A reviewer of this paper remarked that, if we set d = 1, density (26) does not reduce to the form 2 t1.y; ν/T1.αy; ν/, which seems to be the ‘most natural’ univariate form of skew t-density generated by lemma 1 of Azzalini (1985), a forerunner of proposition 1....
[...]
...In connection with the SN distribution, Azzalini (1985) and Azzalini and Capitanio (1999) have highlighted some problematic aspects of the likelihood function....
[...]
...In fact, one could reverse the reasoning and claim that lemma 1 of Azzalini (1985) ‘should’ have been stated in the form of proposition 1 for d = 1; in other words, there is no reason to restrict w.y/ to the linear form αy, especially outside the normal case....
[...]
...The multivariate distribution sketched by Azzalini (1985), section 4, and the multiple-constraint model outlined by Arnold and Beaver (2000a), section 6, has a G which is the product of m (m 1) terms of type Φ.αiyi/ or Φ.αTi y + bi/ respectively....
[...]
2,106 citations
"Distributions generated by perturba..." refers background in this paper
...U|U0 = z/ does not depend on Qz if and only if UÅ is Gaussian; see theorem 4.12 of Fang et al. (1990)....
[...]
...Using results in Fang et al. (1990), pages 82–83, we have c1 f̃ Qz .y2/ = Γ....
[...]
...The proof is identical with that of proposition 4, considering the densities of marginal and conditional distributions of PII, as defined in Fang et al. (1990), pages 89–91....
[...]
...12 of Fang et al. (1990). In this case, the integral in (16) becomes , so that , ! ....
[...]
...For a full treatment of this topic, we refer the reader to Fang et al. (1990)....
[...]
1,478 citations
1,130 citations
"Distributions generated by perturba..." refers background or methods or result in this paper
...Other properties of quadratic forms of SN variables are given by Azzalini & Capitanio (1999) , Genton et al. (2001) and Loperfido (2001)....
[...]
...As a further level of generalisation of the normal distribution, Azzalini & Capitanio (1999, p. 599) have presented a lemma which leads to the construction of a ‘skew elliptical’ density, which is an elliptical density multiplied by a suitable skewing factor, in such a way that the product is still a proper density....
[...]
...while, for the expressions of , we use results given by Azzalini & Capitanio (1999) and by Genton et al. (2001)....
[...]
...The first is to prove that the two forms of skew elliptical densities introduced by Azzalini & Capitanio (1999, p. 599) and by Branco & Dey (2001) are closely connected....
[...]
...Notice that Propositions 8 and 9 of Azzalini & Capitanio (1999) have added conditions on the parameter, but these are not necessary....
[...]
829 citations
"Distributions generated by perturba..." refers methods in this paper
...Alternative proposals of univariate skew t-distributions have been made by Fernández and Steel (1998), constructed similarly to the so-called two-piece normal density, and by Jones (2001), developed by Jones and Faddy (2003), which is based on a suitable transformation of a beta density....
[...]