Skip to main content
Log in

Multivariate Doubly Truncated Moments for a Class of Multivariate Location-Scale Mixture of Elliptical Distributions

  • Published:
Mathematical Methods of Statistics Aims and scope Submit manuscript

Abstract

In this paper, we investigate multivariate doubly truncated moments for a class of multivariate location-scale mixture of elliptical (LSME) distributions. This rich family includes some well-known distributions, such as location-scale mixture of normal, location-scale mixture of Student-\(t\), location-scale mixture of logistic and location-scale mixture of Laplace distributions, as special cases. We first present general formulae for computing the first two moments of the LSME distributions under the double truncation. We then consider a special case for cross moment. As an application, we present the results of multivariate tail conditional expectation (MTCE) for generalized hyperbolic (MGH) distribution.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1

Similar content being viewed by others

REFERENCES

  1. T. Amemiya, ‘‘Multivariate regression and simultaneous equation models when the dependent variables are truncated normal,’’ Econometrica 42, 999–1012 (1974). https://doi.org/10.2307/1914214

    Article  MathSciNet  MATH  Google Scholar 

  2. J. C. Arismendi, ‘‘Multivariate truncated moments,’’ Journal of Multivariate Analysis 117, 41–75 (2013).

    Article  MathSciNet  MATH  Google Scholar 

  3. J. C. Arismendi and S. Broda, ‘‘Multivariate elliptical truncated moments,’’ Journal of Multivariate Analysis 157, 29–44 (2017). https://doi.org/10.1016/j.jmva.2017.02.011

    Article  MathSciNet  MATH  Google Scholar 

  4. N. D. Cameron and R. Thompson, ‘‘Design of multivariate selection experiments to estimate genetic parameters,’’ Theoretical and Applied Genetics 72(4), 466–476 (1986). https://doi.org/10.1017/s0308229600015373

    Article  Google Scholar 

  5. A. Cerioli, A. Farcomeni, and M. Riani, ‘‘Strong consistency and robustness of the forward search estimator of multivariate location and scatter,’’ Journal of Multivariate Analysis 126, 167–183 (2014). https://doi.org/10.1016/j.jmva.2013.12.010

    Article  MathSciNet  MATH  Google Scholar 

  6. K. T. Fang, S. Kotz, and K. W. Ng, Symmetric Multivariate and Related Distributions (Chapman and Hall, New York, 1990).

    Book  MATH  Google Scholar 

  7. C. E. Galarza, T. I. Lin, W. L. Wang, and V. H. Lachos, ‘‘On moments of folded and truncated multivariate Student-\(t\) distributions based on recurrence relations,’’ Metrika 84 (6), 825–850 (2021). https://doi.org/10.1007/s00184-020-00802-1

    Article  MathSciNet  MATH  Google Scholar 

  8. C. E. Galarza, L. A. Matos, L. M. Castro, and V. H. Lachos, ‘‘Moments of the doubly truncated selection elliptical distributions with emphasis on the unified multivariate skew-\(t\) distribution,’’ Journal of Multivariate Analysis 189, 104944 (2022). https://doi.org/10.1016/j.jmva.2021.104944

  9. E. Gómez-Sánchez-Manzano, J. M. Gómez-Villegas, and M. A. Marín, ‘‘Sequences of elliptical distributions and mixtures of normal distributions,’’ Journal of Multivariate Analysis 97, 295–310 (2006). https://doi.org/10.1016/j.jmva.2005.03.008

    Article  MathSciNet  MATH  Google Scholar 

  10. T. Gneiting, ‘‘Normal scale mixtures and dual probability densities,’’ Journal of Statistical Computation and Simulation 59, 375–384 (1997). https://doi.org/10.1080/00949659708811867

    Article  MATH  Google Scholar 

  11. G.A. Hanasusanto, D. Kuhn, S. W. Wallace, and S. Zymler, ‘‘Distributionally robust multi-item newsvendor problems with multimodal demand distributions,’’ Mathematical Programming 152, 1–32 (2014). https://doi.org/10.1007/s10107-014-0776-y

    Article  MathSciNet  MATH  Google Scholar 

  12. H. J. Ho, T. I. Lin, H. Y. Chen, and W. L. Wang, ‘‘Some results on the truncated multivariate \(t\) distribution,’’ Journal of Statistical Planning and Inference 142, 25–40 (2012). https://doi.org/10.1016/j.jspi.2011.06.006

    Article  MathSciNet  MATH  Google Scholar 

  13. K. Ignatieva and Z. Landsman, ‘‘Estimating the tail of loss severity via conditional risk measures for the family of symmetric generalised hyperbolic distributions,’’ Insurance: Mathematics and Economics 65, 172–186 (2015). https://doi.org/10.1016/j.insmatheco.2015.09.007

    Article  MathSciNet  MATH  Google Scholar 

  14. K. Ignatieva and Z. Landsman, ‘‘Conditional tail risk measures for the skewed generalised hyperbolic family,’’ Insurance: Mathematics and Economics 86, 98–114 (2019). https://doi.org/10.1016/j.insmatheco.2019.02.008

    Article  MathSciNet  MATH  Google Scholar 

  15. K. Ignatieva and Z. Landsman, ‘‘A new class of generalised hyper-elliptical distributions and their applications in computing conditional tail risk measures,’’ Insurance: Mathematics and Economics 101, 437–465 (2020). https://doi.org/10.1016/j.insmatheco.2021.08.011

    Article  MATH  Google Scholar 

  16. R. Kan and C. Robotti, ‘‘On moments of folded and truncated multivariate normal distributions,’’ Journal of Computational and Graphical Statistics 26 (4), 930–934 (2017). https://doi.org/10.1080/10618600.2017.1322092

    Article  MathSciNet  Google Scholar 

  17. Z. Landsman, ‘‘On the tail mean-variance optimal portfolio selection,’’ Insurance: Mathematics and Economics 46, 547–553 (2010). https://doi.org/10.1016/j.insmatheco.2010.02.001

    Article  MathSciNet  MATH  Google Scholar 

  18. Z. Landsman, U. Makov, and T. Shushi, ‘‘Multivariate tail conditional expectation for elliptical distributions,’’ Insurance: Mathematics and Economics 70, 216–223 (2016). https://doi.org/10.1016/j.insmatheco.2016.05.017

    Article  MathSciNet  MATH  Google Scholar 

  19. Z. Landsman, U. Makov, and T. Shushi, ‘‘Tail conditional moments for elliptical and log-elliptical distributions,’’ Insurance: Mathematics and Economics 71, 179–188 (2016). https://doi.org/10.1016/j.insmatheco.2016.09.001

    Article  MathSciNet  MATH  Google Scholar 

  20. Z. Landsman and E. A. Valdez, ‘‘Tail conditional expectations for elliptical distributions,’’ North American Actuarial Journal 7, 55–71 (2003). https://doi.org/10.1080/10920277.2003.10596118

    Article  MathSciNet  MATH  Google Scholar 

  21. L. F. Lee, ‘‘On the first and second moments of the truncated multi-normal distribution and a simple estimator,’’ Economics Letters 3, 165–169 (1979). https://doi.org/10.1016/0165-1765(79)90111-3

    Article  MathSciNet  Google Scholar 

  22. Z. Li, J. Luo, and J. Yao, ‘‘Convex bound approximations for sums of random variables under multivariate log-generalized hyperbolic distribution and asymptotic equivalences,’’ Journal of Computational and Applied Mathematics 391, 113459 (2021). https://doi.org/10.1016/j.cam.2021.113459

  23. D. H. D. Lien, ‘‘Moments of truncated bivariate log-normal distributions,’’ Economics Letters 19, 243–247 (1985). https://doi.org/10.1016/0165-1765(85)90029-1

    Article  MathSciNet  MATH  Google Scholar 

  24. B. G. Manjunath and S. Wilhelm, ‘‘Moments calculation for the doubly truncated multivariate normal density,’’ Ssrn Electronic Journal, (2012). https://doi.org/10.2139/ssrn.1472153

    Book  Google Scholar 

  25. A. J. McNeil, R. Frey, and P. Embrechts, Quantitative Risk Management: Concepts, Techniques, and Tools (Princeton University Press, New Jersey, 2005).

    MATH  Google Scholar 

  26. C. E. G. Morales, L. A. Matos, D. K. Dey, and V. H. Lachos, ‘‘On moments of folded and doubly truncated multivariate extended skew-normal distributions,’’ Journal of Computational and Graphical Statistics, (2022). https://doi.org/10.1080/10618600.2021.2000869

  27. H. Ogasawara, ‘‘A non-recursive formula for various moments of the multivariate normal distribution with sectional truncation,’’ Journal of Multivariate Analysis 183, 104729 (2021). https://doi.org/10.1016/j.jmva.2021.104729

  28. R. Roozegar, N. Balakrishnan, and A. Jamalizadeh, ‘‘On moments of doubly truncated multivariate normal meanиCvariance mixture distributions with application to multivariate tail conditional expectation,’’ Journal of Multivariate Analysis 177, 104586 (2020). https://doi.org/10.1016/j.jmva.2019.104586

  29. D. Shi, T. Chen, and L. Shi, ‘‘An event-triggered approach to state estimation with multiple point- and set-valued measurements,’’ Automatica 50 (6), 1641–1648 (2014). https://doi.org/10.1016/j.automatica.2014.04.004

    Article  MathSciNet  MATH  Google Scholar 

  30. G. M. Tallis, ‘‘The moment generating function of the truncated multi-normal distribution,’’ Journal of the Royal Statistical Society. Series B: Methodological 23 (1), 223–229 (1961). https://doi.org/10.1111/j.2517-6161.1961.tb00408

    Article  MathSciNet  MATH  Google Scholar 

  31. M. Weibel, D. Luethi, and W. Breymann, ghyp: Generalized Hyperbolic Distribution and Its Special Cases, r package version 1.6.1 (2020), https://CRAN.R-project.org/package=ghyp.

  32. B. S. Zuo and C. C. Yin, ‘‘Tail conditional risk measures for location-scale mixture of elliptical distributions,’’ Journal of Statistical Computation and Simulation 91 (17), 3653–3677 (2021). https://doi.org/10.1080/00949655.2021.1944142

    Article  MathSciNet  MATH  Google Scholar 

  33. B. S. Zuo and C. C. Yin, ‘‘Multivariate doubly truncated moments for generalized skew-elliptical distributions with application to multivariate tail conditional risk measure,’’ Submitted, (2022).

Download references

ACKNOWLEDGMENTS

The research was supported by the National Natural Science Foundation of China (nos. 12071251, 11571198).

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Chuancun Yin.

Ethics declarations

The authors declare that there is no conflict of interests regarding the publication of this article.

APPENDIX

APPENDIX

In this appendix, we provide all the proofs of the main results, states in Section 3, of this paper.

Proof of Theorem 1. (1) Using the transformation \(\mathbf{Z}=(\theta\boldsymbol{\Sigma})^{-\frac{1}{2}}(\mathbf{Y}-\boldsymbol{\mu}-\theta\boldsymbol{\beta})\), we have

$${\textrm{E}}[{\mathbf{Y}}|{a}<{\mathbf{Y}}\leq{b}]=\frac{1}{\alpha_{1,n}}\int\limits_{a}^{b}{y}f_{\mathbf{Y}}(y)\textrm{d}y=\frac{1}{\alpha_{1,n}}\int\limits_{a}^{b}\int\limits_{\Omega_{\theta}}yf_{\mathbf{Y}|\Theta}(y)f_{\Theta}(\theta){\textrm{d}}\theta{\textrm{d}}{y}$$
$${}=\frac{1}{\alpha_{1,n}}\int\limits_{\Omega_{\theta}}f_{\Theta}(\theta)\int\limits_{a}^{b}{y}\frac{c_{1,n}}{\sqrt{|\theta\boldsymbol{\Sigma}|}},$$
$$g_{n}\left(\frac{1}{2}(y-\mu-\theta\boldsymbol{\beta})^{T}(\theta\boldsymbol{\Sigma})^{-1}(y-\mu-\theta\boldsymbol{\beta})\right)\textrm{d}y\textrm{d}\theta$$
$${}=\frac{1}{\alpha_{1,n}}\!\int\limits_{\Omega_{\theta}}\!f_{\Theta}(\theta)\int\limits_{a^{\prime}}^{b^{\prime}}\!\left(\left(\theta\boldsymbol{\Sigma}\right)^{1/2}\mathbf{z}\!+\boldsymbol{\mu}\!+\theta\boldsymbol{\beta}\right)c_{1,n}g_{n}\left(\frac{1}{2}{\mathbf{z}}^{T}\mathbf{z}\right)\textrm{d}\mathbf{z}\textrm{d}\theta$$
$${}=\frac{1}{\alpha_{1,n}}\int\limits_{\Omega_{\theta}}f_{\Theta}(\theta)\int\limits_{a^{\prime}}^{b^{\prime}}\left(\left(\theta\boldsymbol{\Sigma}\right)^{1/2}\mathbf{z}c_{1,n}g_{n}\{\frac{1}{2}\mathbf{z}^{T}\mathbf{z}\}+(\boldsymbol{\mu}+\theta\boldsymbol{\beta})c_{1,n}g_{n}\left(\frac{1}{2}\mathbf{z}^{T}\mathbf{z}\right)\right)\textrm{d}\mathbf{z}\textrm{d}\theta$$
$${}=\frac{1}{\alpha_{1,n}}\boldsymbol{\Sigma}^{1/2}\int\limits_{\Omega_{\theta}}f_{\Theta}(\theta)\theta^{1/2}\boldsymbol{\delta}\textrm{d}\theta+\int\limits_{\Omega_{\theta}}f_{\Theta}(\theta)(\boldsymbol{\mu}+\theta\boldsymbol{\beta})\textrm{d}\theta$$
$${}=\frac{1}{\alpha_{1,n}}\boldsymbol{\Sigma}^{1/2}\int\limits_{\Omega_{\theta}}f_{\Theta}(\theta)\theta^{1/2}\boldsymbol{\delta}\textrm{d}\theta+\int\limits_{\Omega_{\theta}}f_{\Theta}(\theta)\theta\boldsymbol{\beta}\textrm{d}\theta+\boldsymbol{\mu},$$

where \(a^{\prime}=(a^{\prime}_{1},a^{\prime}_{2},...,a^{\prime}_{n})^{T}=(\theta\boldsymbol{\Sigma})^{-\frac{1}{2}}(a-\boldsymbol{\mu}-\theta\boldsymbol{\beta}),b^{\prime}=(b^{\prime}_{1},b^{\prime}_{2},...,b^{\prime}_{n})^{T}=(\theta\boldsymbol{\Sigma})^{-\frac{1}{2}}(b-\boldsymbol{\mu}-\theta\boldsymbol{\beta})\), and \(\boldsymbol{\delta}=(\delta_{1},\delta_{2},...,\delta_{n})^{T}.\)

Note that

$$\delta_{k}=c_{1,n}\int\limits_{a^{\prime}_{-k}}^{b^{\prime}_{-k}}\int\limits_{a^{\prime}_{k}}^{b^{\prime}_{k}}z_{k}g_{n}\left(\frac{1}{2}\mathbf{z}_{-k}^{T}\mathbf{z}_{-k}+\frac{1}{2}z_{k}^{2}\right)\textrm{d}z_{k}\textrm{d}\mathbf{z}_{-k}=c_{1,n}\int\limits_{a^{\prime}_{-k}}^{b^{\prime}_{-k}}\int\limits_{a^{\prime}_{k}}^{b^{\prime}_{k}}-\partial\overline{G}_{2,n}\left(\frac{1}{2}\mathbf{z}_{-k}^{T}\mathbf{z}_{-k}+\frac{1}{2}z_{k}^{2}\right)\textrm{d}z_{k}\textrm{d}\mathbf{z}_{-k}$$
$${}={c_{1,n}}\int\limits_{a^{\prime}_{-k}}^{b^{\prime}_{-k}}\left(\overline{G}_{2,n}^{a^{\prime}_{k}}\left(\frac{1}{2}\mathbf{z}_{-k}^{T}\mathbf{z}_{-k}+\frac{1}{2}a_{k}^{\prime 2}\right)-\overline{G}_{2,n}^{b^{\prime}_{k}}\left(\frac{1}{2}\mathbf{z}_{-k}^{T}\mathbf{z}_{-k}+\frac{1}{2}b_{k}^{\prime 2}\right)\right)\textrm{d}\mathbf{z}_{-k}$$
$${}=c_{1,n}\left(\frac{\alpha_{2,n-1}^{a^{\prime}_{k}}}{c_{2,n-1}^{a^{\prime}_{k}}}-\frac{\alpha_{2,n-1}^{b^{\prime}_{k}}}{c_{2,n-1}^{b^{\prime}_{k}}}\right).$$

(2) Similarly, using the transformation \(\mathbf{Z}=(\theta\Sigma)^{-\frac{1}{2}}(\mathbf{Y}-\mu-\theta\beta)\), we have

$${\textrm{E}}\left[\mathbf{Y}\mathbf{Y}^{T}|{a}<\mathbf{Y}\leq{b}\right]=\frac{1}{\alpha_{1,n}}\int\limits_{{a}}^{{b}}y{y}^{T}f_{\mathbf{Y}}(y)\textrm{d}{y}=\frac{1}{\alpha_{1,n}}\int\limits_{\Omega_{\theta}}f_{\Theta}(\theta)\int\limits_{{a}}^{{b}}{y}{y}^{T}\frac{c_{1,n}}{\sqrt{|\theta\Sigma|}},$$
$$g_{n}\left(\frac{1}{2}(y-\mu-\theta\beta)^{T}(\theta\Sigma)^{-1}(y-\mu-\theta\beta)\right)\textrm{d}y\textrm{d}\theta$$
$${}=\int\limits_{\Omega_{\theta}}f_{\Theta}(\theta)\left((\theta\Sigma)^{1/2}{\textrm{E}}[\mathbf{ZZ}^{T}|a^{\prime}<\mathbf{Z}\leq b^{\prime}](\theta\Sigma)^{1/2}\right.+(\theta\Sigma)^{1/2}\!{\textrm{E}}[\mathbf{Z}|a^{\prime}<\mathbf{Z}\leq b^{\prime}](\mu+\theta\beta)^{T}$$
$${}+(\mu+\theta\beta){\textrm{E}}[\mathbf{Z}^{T}|a^{\prime}<\mathbf{Z}\leq b^{\prime}](\theta\Sigma)^{1/2}+\left.\theta\mu\beta^{T}+\theta\beta\mu^{T}+\theta^{2}\beta\beta^{T}\right)\textrm{d}\theta+\mu\mu^{T}.$$

For \({\textrm{E}}[\mathbf{Z}\mathbf{Z}^{T}|a^{\prime}<\mathbf{Z}\leq a^{\prime}],\) using the definition of conditional expectation we obtain

$${\textrm{E}}[\mathbf{Z}\mathbf{Z}^{T}|a^{\prime}<\mathbf{Z}\leq a^{\prime}]=\frac{1}{\alpha_{1,n}}\int\limits_{a^{\prime}}^{b^{\prime}}zz^{T}c_{1,n}g_{n}\left(\frac{1}{2}zz^{T}\right)\textrm{d}z$$
$${}=\frac{1}{\alpha_{1,n}}\boldsymbol{\Omega},$$

where \(\Omega=(\Omega_{i,j})_{i,j=1}^{n}.\)

Note that, for \(i=j,\) we have

$$\Omega_{ii}=\int\limits_{a^{\prime}}^{b^{\prime}}z_{i}^{2}c_{1,n}g_{n}\left(\frac{1}{2}{z}{z}^{T}\right)\textrm{d}\mathbf{z}=\int\limits_{a^{\prime}_{-i}}^{b^{\prime}_{-i}}\int\limits_{a^{\prime}_{i}}^{b^{\prime}_{i}}-z_{i}c_{1,n}\partial_{i}\overline{G}_{2,n}\left\{\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}z_{i}^{2}\right\}\textrm{d}z_{i}\textrm{d}\textbf{z}_{-i}$$
$${}=c_{1,n}\int\limits_{a^{\prime}_{-i}}^{b^{\prime}_{-i}}\left(a^{\prime}_{i}\overline{G}_{2,n}^{a^{\prime}_{i}}\left(\frac{1}{2}z_{-i}^{T}z_{-i}+\frac{1}{2}a_{i}^{\prime 2}\right)-b^{\prime}_{i}\overline{G}_{2,n}^{b^{\prime}_{i}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}b_{i}^{\prime 2}\right)\right.$$
$${}+\left.\int\limits_{a^{\prime}}^{b^{\prime}}\overline{G}_{2,n}\left(\frac{1}{2}zz^{T}\right)\textrm{d}z_{i}\right)\textrm{d}z_{-i}$$
$${}=\int\limits_{a^{\prime}_{-i}}^{b^{\prime}_{-i}}\left(a^{\prime}_{i}\frac{c_{1,n}}{c_{2,n-1}^{a^{\prime}_{i}}}c_{2,n-1}^{a^{\prime}_{i}}\overline{G}_{2,n}^{a^{\prime}_{i}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}a_{i}^{\prime 2}\right)\right.$$
$${}-\left.b^{\prime}_{i}\frac{c_{1,n}}{c_{2,n-1}^{b^{\prime}_{i}}}c_{2,n-1}^{b^{\prime}_{i}}\overline{G}_{2,n}^{b^{\prime}_{i}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}b_{i}^{\prime 2}\right)\right)\textrm{d}z_{-i}+\int\limits_{a^{\prime}}^{b^{\prime}}c_{1n}\overline{G}_{2,n}\left(\frac{1}{2}zz^{T}\right)\textrm{d}z$$
$${}=\frac{c_{1,n}}{c_{2,n-1}^{a^{\prime}_{i}}}a^{\prime}_{i}\alpha_{2,n-1}^{a^{\prime}_{i}}-\frac{c_{1,n}}{c_{2,n-1}^{b^{\prime}_{i}}}b^{\prime}_{i}\alpha_{2,n-1}^{b^{\prime}_{i}}+\frac{c_{1,n}}{c_{2,n}}\alpha_{2,n},$$

while \(i\neq j,\)

$$\Omega_{ij}=\int\limits_{a^{\prime}}^{b^{\prime}}z_{i}z_{j}c_{1,n}g_{n}\left(\frac{1}{2}{z}{z}^{T}\right)\textrm{d}{z}=\int\limits_{a^{\prime}_{-i}}^{b^{\prime}_{-i}}\int\limits_{a^{\prime}_{i}}^{b^{\prime}_{i}}-z_{j}c_{1,n}\partial_{i}\overline{G}_{2,n}\left(\frac{1}{2}\mathbf{z}_{-i}\mathbf{z}_{-i}^{T}+\frac{1}{2}z_{i}^{2}\right)\textrm{d}z_{i}\textrm{d}{z}_{-i}$$
$${}=\int\limits_{a^{\prime}_{-i}}^{b^{\prime}_{-i}}\!\left(z_{j}c_{1,n}\overline{G}_{2,n-1}^{a^{\prime}_{i}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\!\frac{1}{2}a_{i}^{\prime 2}\right)\right.-\left.z_{j}c_{1,n}\overline{G}_{2,n-1}^{b^{\prime}_{i}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}b_{i}^{\prime 2}\right)\right)\textrm{d}z_{-i}$$
$${}=c_{1,n}\int\limits_{a^{\prime}_{-i,-j}}^{b^{\prime}_{-i,-j}}\int\limits_{a^{\prime}_{j}}^{b^{\prime}_{j}}\left(z_{j}\overline{G}_{2,n-1}^{a^{\prime}_{i}}\left(\frac{1}{2}z_{-i,-j}z_{-i,-j}^{T}+\frac{1}{2}z_{j}^{2}+\frac{1}{2}a_{i}^{\prime 2}\right)\right.$$
$${}-\left.z_{j}\overline{G}_{2,n-1}^{b^{\prime}_{i}}\left(\frac{1}{2}z_{-i,-j}z_{-i,-j}^{T}+\frac{1}{2}z_{j}^{2}+\frac{1}{2}b_{i}^{\prime 2}\right)\right)\textrm{d}z_{j}\textrm{d}z_{-i,-j}$$
$${}=c_{1,n}\int\limits_{a^{\prime}_{-i,-j}}^{b^{\prime}_{-i,-j}}\int\limits_{a^{\prime}_{j}}^{b^{\prime}_{j}}\left(-\partial_{j}\overline{G}_{3,n-1}^{a^{\prime}_{i}}\left(\frac{1}{2}\textbf{z}_{-i,-j}\textbf{z}_{-i,-j}^{T}+\frac{1}{2}z_{j}^{2}+\frac{1}{2}a_{i}^{\prime 2}\right)\right.$$
$${}+\left.\partial_{j}\overline{G}_{3,n-1}^{b^{\prime}_{i}}\left(\frac{1}{2}z_{-i,-j}z_{-i,-j}^{T}+\frac{1}{2}z_{j}^{2}+\frac{1}{2}b_{i}^{\prime 2}\right)\right)\textrm{d}z_{j}\textrm{d}z_{-i,-j}$$
$${}=c_{1,n}\int\limits_{a^{\prime}_{-i,-j}}^{b^{\prime}_{-i,-j}}\left(\overline{G}_{3,n-2}^{a^{\prime}_{i},a^{\prime}_{j}}\left(\frac{1}{2}z_{-i,-j}z_{-i,-j}^{T}+\frac{1}{2}a_{j}^{\prime 2}+\frac{1}{2}a_{i}^{\prime 2}\right)\right.$$
$${}-\left.\overline{G}_{3,n-2}^{a^{\prime}_{i},b^{\prime}_{i}}\left(\frac{1}{2}z_{-i,-j}z_{-i,-j}^{T}+\frac{1}{2}a_{i}^{\prime 2}+\frac{1}{2}b_{i}^{\prime 2}\right)\right.+\left.\overline{G}_{3,n-2}^{b^{\prime}_{i},b^{\prime}_{j}}\left(\frac{1}{2}z_{-i,-j}z_{-i,-j}^{T}+\frac{1}{2}b_{j}^{\prime 2}+\frac{1}{2}b_{i}^{\prime 2}\right)\right.$$
$${}-\left.\overline{G}_{3,n-2}^{a^{\prime}_{j},b^{\prime}_{i}}\left(\frac{1}{2}z_{-i,-j}z_{-i,-j}^{T}+\frac{1}{2}a_{j}^{\prime 2}+\frac{1}{2}b_{i}^{\prime 2}\right)\right)\textrm{d}z_{-i,-j}$$
$${}=\frac{{c_{1,n}}}{c_{3,n-2}^{a^{\prime}_{j}a^{\prime}_{i}}}\alpha_{3,n-2}^{a^{\prime}_{j}a^{\prime}_{i}}-\frac{{c_{1,n}}}{c_{3,n-2}^{b^{\prime}_{j}a^{\prime}_{i}}}\alpha_{3,n-2}^{b^{\prime}_{j}a^{\prime}_{i}}+\frac{{c_{1,n}}}{c_{3,n-2}^{b^{\prime}_{j}b^{\prime}_{i}}}\alpha_{3,n-2}^{b^{\prime}_{j}b^{\prime}_{i}}-\frac{{c_{1,n}}}{c_{3,n-2}^{a^{\prime}_{j}b^{\prime}_{i}}}\alpha_{3,n-2}^{a^{\prime}_{j}b^{\prime}_{i}},$$

where we have used integration by parts in the third and fifth equalities.

As for \({\textrm{E}}[\mathbf{Z}|a^{\prime}<\mathbf{Z}\leq b^{\prime}],\) using (1) we directly obtain

$${\textrm{E}}[\mathbf{Z}|a^{\prime}<\mathbf{Z}\leq b^{\prime}]=\frac{\boldsymbol{\delta}}{\alpha_{1,n}}.$$

Consequently, we obtain (7), ending the proof of (2).

Proof of Remark 1.

$${\textrm{E}}[\mathbf{YY^{T}Y}|a<\mathbf{Y}\leq b]=\frac{1}{\alpha_{1,n}}\int\limits_{\Omega}f(\theta)\int\limits_{a^{\prime}}^{b^{\prime}}\left[\left(\theta\Sigma\right)^{1/2}\mathbf{z}+\mu+\theta\beta\right]$$
$$\left[\mathbf{z}^{T}\left(\theta\Sigma\right)^{1/2}+\mu^{T}+\theta\beta^{T}\right]\left[\left(\theta\Sigma\right)^{1/2}\mathbf{z}+\mu+\theta\beta\right]\textrm{d}\mathbf{z}\textrm{d}\theta$$
$${}=\frac{1}{\alpha_{1,n}}\int\limits_{\Omega}f(\theta)\int\limits_{a^{\prime}}^{b^{\prime}}\left[(\theta\Sigma)^{1/2}\mathbf{z}\mathbf{z}^{T}(\theta\Sigma)^{1/2}(\theta\Sigma)^{1/2}\mathbf{z}\right.$$
$${}+(\theta\Sigma)^{1/2}\mathbf{z}\mathbf{z}^{T}(\theta\Sigma)^{1/2}(\mu+\theta\beta)$$
$${}+(\theta\Sigma)^{1/2}\mathbf{z}\mu^{T}\left((\theta\Sigma)^{1/2}\mathbf{z}+\mu+\theta\beta\right)+(\theta\Sigma)^{1/2}\mathbf{z}(\theta\beta)^{T}\left((\theta\Sigma)^{1/2}\mathbf{z}+\mu+\theta\beta\right)$$
$${}+\mu\mathbf{z}^{T}(\theta\Sigma)^{1/2}\left((\theta\Sigma)^{1/2}\mathbf{z}+\mu+\theta\beta\right)+\mu\mu^{T}\left((\theta\Sigma)^{1/2}\mathbf{z}+\mu+\theta\beta\right)$$
$${}+\theta\mu\beta^{T}\left((\theta\Sigma)^{1/2}\mathbf{z}+\mu+\theta\beta\right)+\theta\beta\mathbf{z}^{T}(\theta\Sigma)^{1/2}\left((\theta\Sigma)^{1/2}\mathbf{z}+\mu+\theta\beta\right)$$
$${}+\theta\beta\mu^{T}\left((\theta\Sigma)^{1/2}\mathbf{z}+\mu+\theta\beta\right)+\left.\theta^{2}\beta\beta^{T}\left((\theta\Sigma)^{1/2}\mathbf{z}+\mu+\theta\beta\right)\right]\textrm{d}\mathbf{z}\textrm{d}\theta$$
$${}=\int\limits_{\Omega_{\theta}}f_{\Theta}(\theta)\left[\theta^{3/2}\Sigma^{1/2}\textrm{E}[\mathbf{Z}\mathbf{Z}^{T}\Sigma\mathbf{Z}|a^{\prime}<\mathbf{Z}\leq b^{\prime}]\right.+(\theta\Sigma)^{1/2}\textrm{E}[\mathbf{Z}\mathbf{Z}^{T}|a^{\prime}<\mathbf{Z}\leq b^{\prime}](\theta\Sigma)^{1/2}(\mu+\theta\beta)$$
$${}+(\theta\Sigma)^{1/2}\textrm{E}[\mathbf{Z}\mu^{T}(\theta\Sigma)^{1/2}\mathbf{Z}|a^{\prime}<\mathbf{Z}\leq b^{\prime}]+(\theta\Sigma)^{1/2}\textrm{E}[\mathbf{Z}(\theta\beta)^{T}(\theta\Sigma)^{1/2}\mathbf{Z}|a^{\prime}<\mathbf{Z}\leq b^{\prime}]$$
$${}+\theta(\mu+\beta)\textrm{E}[\mathbf{Z}^{T}\Sigma\mathbf{Z}|a^{\prime}<\mathbf{Z}\leq b^{\prime}]+(\theta\Sigma)^{1/2}\textrm{E}[\mathbf{Z}|a^{\prime}<\mathbf{Z}\leq b^{\prime}]\left(\mu^{T}+(\theta\beta)^{T}\right)(\mu+\theta\beta)$$
$${}+\left(\mu\mu^{T}+\theta\mu\beta^{T}+\theta\beta\mu^{T}+\theta^{2}\beta\beta^{T}\right)(\theta\Sigma)^{1/2}\textrm{E}[\mathbf{Z}|a^{\prime}<\mathbf{Z}\leq b^{\prime}]$$
$${}+(\mu+\theta\beta)\textrm{E}[\mathbf{Z}^{T}|a^{\prime}<\mathbf{Z}\leq b^{\prime}](\theta\Sigma)^{1/2}(\mu+\theta\beta)$$
$${}+\left.\left(\mu\mu^{T}+\theta\mu\beta^{T}+\theta\beta\mu^{T}+\theta^{2}\beta\beta^{T}\right)(\mu+\theta\beta)\right]\textrm{d}\theta.$$

As for \(\textrm{E}[\mathbf{Z}\mathbf{Z}^{T}\Sigma\mathbf{Z}|a^{\prime}<\mathbf{Z}\leq b^{\prime}]\), its result is an \(n\)-dimensional vector formed by a linear combination of the following three formulas.For \(i\neq j\neq k,\)

$$\int\limits_{a^{\prime}}^{b^{\prime}}z_{i}z_{j}z_{k}c_{1,n}g_{n}\left(\frac{1}{2}\mathbf{z}\mathbf{z}^{T}\right)\textrm{d}\mathbf{z}$$
$${}=c_{1,n}\int\limits_{a_{-i,-j}^{\prime}}^{b_{-i,-j}^{\prime}}\int\limits_{a_{j}}^{b_{j}^{\prime}}\left(-z_{k}\partial_{j}\overline{G}_{3,n-1}^{a_{i}^{\prime}}\left(\frac{1}{2}\mathbf{z}_{-i,-j}\mathbf{z}_{-i,-j}^{T}+\frac{1}{2}z_{j}^{2}+\frac{1}{2}a_{i}^{\prime 2}\right)\right.$$
$${}+\left.z_{k}\partial_{j}\overline{G}_{3,n-1}^{b_{i}^{\prime}}\left(\frac{1}{2}\mathbf{z}_{-i,-j}\mathbf{z}_{-i,-j}^{T}+\frac{1}{2}z_{j}^{2}+\frac{1}{2}b_{i}^{\prime 2}\right)\right)\textrm{d}\textrm{z}_{j}\textrm{d}\mathbf{z}_{-i,-j}$$
$${}=\frac{c_{1,n}}{c_{4,n-3}^{a_{i}^{\prime}a_{j}^{\prime}a_{k}^{\prime}}}\alpha_{4,n-3}^{a_{i}^{\prime}a_{j}^{\prime}a_{k}^{\prime}}-\frac{c_{1,n}}{c_{4,n-3}^{a_{i}^{\prime}a_{j}^{\prime}b_{k}^{\prime}}}\alpha_{4,n-3}^{a_{i}^{\prime}a_{j}^{\prime}b_{k}^{\prime}}+\frac{c_{1,n}}{c_{4,n-3}^{a_{i}^{\prime}b_{j}^{\prime}b_{k}^{\prime}}}\alpha_{4,n-3}^{a_{i}^{\prime}b_{j}^{\prime}b_{k}^{\prime}}-\frac{c_{1,n}}{c_{4,n-3}^{a_{i}^{\prime}b_{j}^{\prime}a_{k}^{\prime}}}\alpha_{4,n-3}^{a_{i}^{\prime}b_{j}^{\prime}a_{k}^{\prime}},$$

while \(i=k\neq j\),

$$\int\limits_{a^{\prime}}^{b^{\prime}}z_{i}^{2}z_{j}c_{1,n}g_{n}\left(\frac{1}{2}\mathbf{z}z^{T}\right)\textrm{d}\mathbf{z}$$
$${}=c_{1,n}\int\limits_{a_{-i,-j}^{\prime}}^{b_{-i,-j}^{\prime}}\int\limits_{a_{j}^{\prime}}^{b_{j}^{\prime}}\left(a_{i}^{\prime}z_{j}\overline{G}_{2,n-1}^{a_{i}^{\prime}}\left(\frac{1}{2}\mathbf{z}_{-i}\mathbf{z}_{-i}^{T}+\frac{1}{2}a_{i}^{\prime 2}\right)\right.$$
$${}-\left.b_{i}^{\prime}z_{j}\overline{G}_{2,n-1}^{b_{i}^{\prime}}\left(\frac{1}{2}\mathbf{z}_{-i}\mathbf{z}_{-i}^{T}+\frac{1}{2}b_{i}^{\prime 2}\right)\right)\textrm{d}z_{-j}\textrm{d}\mathbf{z}_{-i,-j}+c_{1,n}\int\limits_{a^{\prime}}^{b^{\prime}}z_{j}\overline{G}_{2,n}\left(\frac{1}{2}\mathbf{zz^{T}}\right)\textrm{d}\mathbf{z}$$
$${}=a_{i}^{\prime}\frac{c_{1,n}}{c_{3,n-2}^{a_{i}^{\prime}a_{j}^{\prime}}}\alpha_{3,n-2}^{a_{i}^{\prime}a_{j}^{\prime}}-b_{i}^{\prime}\frac{c_{1,n}}{c_{3,n-2}^{a_{i}^{\prime}a_{j}^{\prime}}}\alpha_{3,n-2}^{a_{i}^{\prime}a_{j}^{\prime}}+b_{i}^{\prime}\frac{c_{1,n}}{c_{3,n-2}^{b_{i}^{\prime}b_{j}^{\prime}}}\alpha_{3,n-2}^{b_{i}^{\prime}b_{j}^{\prime}}$$
$${}-b_{i}^{\prime}\frac{c_{1,n}}{c_{3,n-2}^{b_{i}^{\prime}a_{j}^{\prime}}}\alpha_{3,n-2}^{b_{i}^{\prime}a_{j}^{\prime}}+\frac{c_{1,n}}{c_{3,n-1}^{a_{j}^{\prime}}}\alpha_{3,n-1}^{a_{j}^{\prime}}-\frac{c_{1,n}}{c_{3,n-1}^{b_{j}^{\prime}}}\alpha_{3,n-1}^{b_{j}^{\prime}},$$

while \(i=j=k\),

$$\int\limits_{a^{\prime}}^{b^{\prime}}z_{i}^{3}c_{1,n}g_{n}\left(\frac{1}{2}\mathbf{zz^{T}}\right)\textrm{d}\mathbf{z}=c_{1,n}\int\limits_{a_{-i}^{\prime}}^{b_{-i}^{\prime}}\left(a_{i}^{\prime 2}\overline{G}_{2,n-1}^{a_{i}^{\prime}}\left(\frac{1}{2}\mathbf{z}_{-i}\mathbf{z}_{-i}^{T}+\frac{1}{2}a_{i}^{\prime 2}\right){}\right.$$
$${}\left.-b_{i}^{\prime 2}\overline{G}_{2,n-1}^{b_{i}^{\prime}}\left(\frac{1}{2}\mathbf{z}_{-i}\mathbf{z}_{-i}^{T}+\frac{1}{2}b_{i}^{\prime 2}\right)\right)\textrm{d}\mathbf{z}_{-i}+c_{1,n}\int\limits_{a^{\prime}}^{b^{\prime}}2z_{i}\overline{G}_{2,n}\left(\frac{1}{2}\mathbf{zz^{T}}\right)\textrm{d}\mathbf{z}.$$

Proof of Theorem 2. Using the transformation \(\mathbf{Z}=(\theta\Sigma)^{1/2}(\mathbf{Y}-\mu_{d})\), we have

$${\textrm{E}}_{\alpha_{1,n}}(\widetilde{\mathbf{Y}}_{d}^{\mathbf{k}})=\frac{1}{\alpha_{1,n}}\int\limits_{a_{d}}^{b_{d}}{\widetilde{y}}_{d}^{\mathbf{k}}f_{\mathbf{Y}}(y)\textrm{d}y=\frac{1}{\alpha_{1,n}}\int\limits_{\Omega_{\theta}}f_{\Theta}(\theta)\int\limits_{a_{d}}^{b_{d}}(\widetilde{y}_{d})^{\mathbf{k}}f_{\mathbf{Y}|\Theta}(y|\Theta)\textrm{d}y\textrm{d}\theta$$
$${}=\frac{1}{\alpha_{1,n}}\!\int\limits_{\Omega_{\theta}}\!f_{\Theta}(\theta)\!\int\limits_{a_{d}}^{b_{d}}\!(\widetilde{y}_{d})^{\mathbf{k}}\!\frac{c_{1,n}}{\sqrt{\theta\Sigma}}\!g_{n}\left(\frac{1}{2}({y\!-\mu_{d}})^{T}(\theta\Sigma)^{-1}({y\!-\mu_{d}})\right)\textrm{d}y\textrm{d}\theta$$
$${}=\frac{1}{\alpha_{1,n}}\int\limits_{\Omega_{\theta}}f_{\Theta}(\theta)\int\limits_{a^{\prime}_{d}}^{b^{\prime}_{d}}\left((\theta\Sigma)^{1/2}z+\mu+\theta\beta\right)^{\mathbf{k}}c_{1,n}g_{n}\left(\frac{1}{2}zz^{T}\right)\textrm{d}z\textrm{d}\theta$$
$${}=\frac{1}{\alpha_{1,n}}C_{k}^{l}\int\limits_{\Omega_{\theta}}f_{\Theta}(\theta)(\theta\Sigma)^{\frac{l}{2}}(\mu+\theta\beta)^{k-l}\int\limits_{a^{\prime}_{d}}^{b^{\prime}_{d}}z^{l}c_{1,n}g_{n}\left(\frac{1}{2}zz^{T}\right)\textrm{d}z\textrm{d}\theta$$
$${}=\frac{1}{\alpha_{1,n}}C_{k}^{l}\int\limits_{\Omega_{\theta}}f_{\Theta}(\theta)(\theta\Sigma)^{\frac{l}{2}}(\mu+\theta\beta)^{k-l}\Lambda\textrm{d}\theta,$$

where \(\Lambda=(\Lambda_{1},\Lambda_{2},...,\Lambda_{n})^{T}\).

Note that

$$\Lambda_{i}=\int\limits_{a^{\prime}_{d}}^{b^{\prime}_{d}}z_{i}^{l}c_{1,n}g_{n}\left(\frac{1}{2}zz^{T}\right)\textrm{d}z=c_{1,n}\int\limits_{a^{\prime}_{d-i}}^{b^{\prime}_{d-i}}\int\limits_{a^{\prime}_{di}}^{b^{\prime}_{di}}-z_{i}^{l-1}\partial_{i}\overline{G}_{2,n}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}z_{i}^{2}\right)\textrm{d}z_{i}\textrm{d}z_{-i}$$
$${}=c_{1,n}\int\limits_{a^{\prime}_{d-i}}^{b^{\prime}_{d-i}}\left[\left({a^{\prime}_{di}}^{l-1}\overline{G}_{2,n}^{a^{\prime}_{di}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}a_{di}^{\prime 2}\right)-{b^{\prime}_{di}}^{l-1}\overline{G}_{2,n}^{b^{\prime}_{di}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}b_{di}^{\prime 2}\right)\right)\right.$$
$${}+\left.(l-1)\int\limits_{a^{\prime}_{di}}^{b^{\prime}_{di}}z_{i}^{l-2}\overline{G}_{2,n}\left(\frac{1}{2}zz^{T}\right)\textrm{d}z_{i}\right]\textrm{d}z_{-i}$$
$${}=c_{1,n}\int\limits_{a^{\prime}_{d-i}}^{b^{\prime}_{d-i}}\left[\left({a^{\prime}_{di}}^{l-1}\overline{G}_{2,n}^{a^{\prime}_{di}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}a_{di}^{\prime 2}\right)-{b^{\prime}_{di}}^{l-1}\overline{G}_{2,n}^{b^{\prime}_{di}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}b_{di}^{\prime 2}\right)\right)\right.$$
$${}+(l-1)\left({a^{\prime}_{di}}^{l-3}\overline{G}_{3,n}^{a^{\prime}_{di}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}a_{di}^{\prime 2}\right)-{b^{\prime}_{di}}^{l-3}\overline{G}_{3,n}^{b^{\prime}_{di}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}b_{di}^{\prime 2}\right)\right)$$
$${}+\left.(l-1)(l-3)\int\limits_{a^{\prime}_{di}}^{b^{\prime}_{di}}z_{i}^{l-4}\overline{G}_{3,n}\left(\frac{1}{2}zz^{T}\right)\textrm{d}z_{i}\right]\textrm{d}z_{-i}.$$

By summarizing, the result can be divided into odd and even cases.

When \(l\) is odd,

$$\Lambda_{i}=c_{1,n}\int\limits_{a^{\prime}_{d-i}}^{b^{\prime}_{d-i}}\left[a_{di}^{\prime l-1}\overline{G}_{2,n-1}^{a^{\prime}_{di}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}a_{di}^{\prime 2}\right)-b_{di}^{\prime l-1}\overline{G}_{2,n-1}^{b^{\prime}_{di}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}b_{di}^{\prime 2}\right)\right.$$
$${}+(l-1)\left(a_{di}^{\prime l-3}\overline{G}_{3,n-1}^{a^{\prime}_{di}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}a_{di}^{\prime 2}\right)-b_{di}^{\prime l-1}\overline{G}_{3,n-1}^{b^{\prime}_{di}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}b_{di}^{\prime 2}\right)\right)$$
$${}+...$$
$${}+(l-1)(l-3)...\left(l-(2m-3)\right)\left(a_{di}^{\prime(l-(2m-1))}\overline{G}_{m+1,n-1}^{a^{\prime}_{di}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}a_{di}^{\prime 2}\right)\right.$$
$${}-\left.b_{di}^{\prime(l-(2m-1))}\overline{G}_{m+1,n-1}^{b^{\prime}_{di}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}b_{di}^{\prime 2}\right)\right)$$
$${}+...$$
$${}+(l-1)(l-3)...1\left(\overline{G}_{\frac{l+3}{2},n-1}^{a_{di}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}a_{di}^{\prime 2}\right)\right.$$
$${}-\left.\left.\overline{G}_{\frac{l+3}{2},n-1}^{b_{di}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}b_{di}^{\prime 2}\right)\right)\right]\textrm{d}z_{-i}$$
$${}=\left(a_{di}^{\prime l-1}\frac{c_{1,n}}{c_{2,n-1}^{a^{\prime}_{di}}}\alpha_{2,n-1}^{a^{\prime}_{di}}-b_{di}^{\prime l-1}\frac{c_{1,n}}{c_{2,n-1}^{b^{\prime}_{di}}}\alpha_{2,n-1}^{b^{\prime}_{di}}\right)$$
$${}+(l-1)\left(a_{di}^{\prime l-3}\frac{c_{1,n}}{c_{3,n-1}^{a^{\prime}_{di}}}\alpha_{3,n-2}^{a^{\prime}_{di}}-b_{di}^{\prime l-3}\frac{c_{1,n}}{c_{3,n-1}^{b^{\prime}_{di}}}\alpha_{3,n-1}^{b^{\prime}_{di}}\right)$$
$${}+...+(l-1)(l-3)...(l-(2m-3))$$
$${}\times\left(a_{di}^{\prime l-(2m-1)}\frac{c_{1,n}}{c_{m+1,n-1}^{a^{\prime}_{di}}}\alpha_{m+1,n-1}^{a^{\prime}_{di}}-b_{di}^{\prime l-(2m-1)}\frac{c_{1,n}}{c_{m+1,n-1}^{b^{\prime}_{di}}}\alpha_{m+1,n-1}^{b^{\prime}_{di}}\right)+...$$
$${}+(l-1)(l-3)...1\left(\frac{c_{1,n}}{c_{\frac{l+3}{2},n-1}^{a^{\prime}_{di}}}\alpha_{\frac{l+3}{2},n-1}^{a^{\prime}_{di}}-\frac{c_{1,n}}{c_{\frac{l+3}{2},n-1}^{b^{\prime}_{di}}}\alpha_{\frac{l+3}{2},n-1}^{b^{\prime}_{di}}\right).$$

When \(l\) is even,

$$\Lambda_{i}=c_{1,n}\int\limits_{a^{\prime}_{d-i}}^{b^{\prime}_{d-i}}\left[a_{di}^{\prime l-1}\overline{G}_{2,n-1}^{a^{\prime}_{di}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}a_{di}^{\prime 2}\right)-b_{di}^{\prime l-1}\overline{G}_{2,n-1}^{b^{\prime}_{di}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}b_{di}^{\prime 2}\right)\right.$$
$${}+(l-1)\left(a_{di}^{\prime l-3}\overline{G}_{3,n-1}^{a^{\prime}_{di}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}a_{di}^{\prime 2}\right)\right.$$
$${}-\left.b_{di}^{\prime l-1}\overline{G}_{2,n-1}^{b^{\prime}_{di}}\overline{G}_{2,n-1}^{b^{\prime}_{di}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}b_{di}^{\prime 2}\right)\right)$$
$${}+...$$
$${}+(l-1)(l-3)...\left(l-(2m-3)\right)\left(a_{di}^{\prime(l-(2m-1))}\overline{G}_{m+1,n-1}^{a^{\prime}_{di}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}a_{di}^{\prime 2}\right)\right.$$
$${}-\left.b_{di}^{\prime(l-(2m-1))}\overline{G}_{m+1,n-1}^{b^{\prime}_{di}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}b_{di}^{\prime 2}\right)\right)$$
$${}+...$$
$${}+(l-1)(l-3)...3\left(\overline{G}_{\frac{l}{2}+1,n-1}^{a_{di}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}a_{di}^{\prime 2}\right)\right.$$
$${}-\left.\left.\overline{G}_{\frac{l}{2}+1,n-1}^{b_{di}}\left(\frac{1}{2}z_{-i}z_{-i}^{T}+\frac{1}{2}b_{di}^{\prime 2}\right)\right)\right]\textrm{d}z_{-i}+(l-1)(l-2)...1c_{1,n}\frac{\alpha_{\frac{l}{2}+1,n}}{c_{\frac{l}{2}+1,n}}$$
$${}=\left(a_{di}^{\prime l-1}\frac{c_{1,n}}{c_{2,n-1}^{a^{\prime}_{di}}}\alpha_{2,n-1}^{a^{\prime}_{di}}-b_{di}^{\prime l-1}\frac{c_{1,n}}{c_{2,n-1}^{b^{\prime}_{di}}}\alpha_{2,n-1}^{b^{\prime}_{di}}\right)$$
$${}+(l-1)\left(a_{di}^{\prime l-3}\frac{c_{1,n}}{c_{3,n-1}^{a^{\prime}_{di}}}\alpha_{3,n-2}^{a^{\prime}_{di}}-b_{di}^{\prime l-3}\frac{c_{1,n}}{c_{3,n-1}^{b^{\prime}_{di}}}\alpha_{3,n-1}^{b^{\prime}_{di}}\right)$$
$${}+...+(l-1)(l-3)...(l-(2m-3))$$
$${}\times\left(a_{di}^{\prime l-(2m-1)}\frac{c_{1,n}}{c_{m+1,n-1}^{a^{\prime}_{di}}}\alpha_{m+1,n-1}^{a^{\prime}_{di}}-b_{di}^{\prime l-(2m-1)}\frac{c_{1,n}}{c_{m+1,n-1}^{b^{\prime}_{di}}}\alpha_{m+1,n-1}^{b^{\prime}_{di}}\right)+...$$
$${}+(l-1)(l-3)...3\left(\frac{c_{1,n}}{c_{\frac{l}{2}+1,n-1}^{a^{\prime}_{di}}}\alpha_{\frac{l}{2}+1,n-1}^{a^{\prime}_{di}}-\frac{c_{1,n}}{c_{\frac{l}{2}+1,n-1}^{b^{\prime}_{di}}}\alpha_{\frac{l}{2}+1,n-1}^{b^{\prime}_{di}}\right)$$
$${}+(l-1)(l-2)...1c_{1,n}\frac{\alpha_{\frac{l}{2}+1,n}}{c_{\frac{l}{2}+1,n}}{.}$$

This completes the proof of Theorem 2.

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Han, X., Yin, C. Multivariate Doubly Truncated Moments for a Class of Multivariate Location-Scale Mixture of Elliptical Distributions. Math. Meth. Stat. 32, 155–175 (2023). https://doi.org/10.3103/S1066530723030043

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.3103/S1066530723030043

Keywords:

Navigation