Abstract
In this paper it is shown how a general two-sided orthant probability for a quadrivariate normal distribution can be evaluated by a one-dimensional numerical integral calculation. The quadrivariate normal distribution can have any covariance matrix and any mean vector. This affords a practical and efficient method for the calculation of these probabilities which is superior to simulation methods. The implementation of the algorithm is discussed, and some examples of its performance are provided.
Similar content being viewed by others
References
Abrahamson IG (1964) Orthant probabilities for the quadrivariate normal distribution. Ann Math Stat 35(4): 1685–1703
Bretz F, Genz A, Hothorn LA (2001) On the numerical availability of multiple comparison procedures. Biom J 43: 645–656
Cheng MC (1969) The orthant probabilities of four Gaussian variates. Ann Math Stat 40(1): 152–161
Gehrlein WV (1979) A representation for quadrivariate normal positive orthant probabilities. Commun Stat B-Simul 8(4): 349–358
Genz A, Bretz F (2002) Methods for the computation of multivariate t-probabilities. J Comput Graph Stat 11: 950–971
Genz A, Bretz F (2009) Computation of multivariate normal and t probabilities. Springer, Heidelberg
Gibson GJ, Glasbey CA, Elston DA (1994) Monte-Carlo evaluation of multivariate normal integrals and sensitivity to variate ordering. Advances in numerical methods and applications, Sofia
Hayter AJ (2006) Recursive integration methodologies with statistical applications. J Stat Plan Infer 136: 2284–2296
Hayter AJ (2011) Recursive integration methodologies for the evaluation of multivariate normal probabilities. J Stat Theory Prac (To appear)
McFadden JA (1960) Two expansions for the quadrivariate normal integral. Biometrika 47: 325–333
Miwa T, Hayter AJ, Kuriki S (2003) The evaluation of general non-centred orthant probabilities. J Roy Stat Soc B 65: 223–234
Patel JK, Read CB (1996) Handbook of the normal distribution. Marcel Dekker, New York
Shampine LF, Allen RC, Pruess S (1997) Fundamentals of numerical computing. Wiley, New York
Schervish MJ (1984) Multivariate normal probabilities with error bound. J Roy Stat Soc C-App 33(1): 81–94
Somerville PN (1998) Numerical computation of multivariate normal and multivariate t probabilities over convex regions. J Comput Graph Stat 7: 529–544
Sondhi MM (1961) A note on the quadrivariate normal integral. Biometrika 48: 201–203
Thisted RA (1988) Elements of statistical computing. Chapman and Hall, New York
Uusipaikka E (1985) Exact simultaneous confidence-intervals for multiple comparisons among 3 or 4 mean-values. J Amer Stat Assoc 80: 196–201
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Hayter, A.J., Lin, Y. The evaluation of two-sided orthant probabilities for a quadrivariate normal distribution. Comput Stat 27, 459–471 (2012). https://doi.org/10.1007/s00180-011-0267-z
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00180-011-0267-z