Summary
For an arbitrary 2n-dimensional centered multivariate normal random vectorX=(X 1, ...,X 2n ) the distribution of\(Z: = \mathop \sum \limits_{i = 1}^n X_i X_{n + 1} \) is computed. It turns out to be a convolution of a collection ofx 21 -distributions with scale parameters depending on the covariance matrix ofX. Forn=1, the case of the bivariate normal distribution, the characteristic function ofZ is explicitly given also in the case of non-centeredX.
Similar content being viewed by others
Literaturverzeichnis
Erdélyi, H. (Hrsg.) mitW. Magnus, F. Oberhettinger undF. Tricomi: Tables of integral transforms. Vol. 1. New York-Toronto-London 1954.
Miller, K.S.: Multidimensional Gaussian distributions. New York-London-Sydney 1964.
—: Some multivariate density functions of products of Gaussian variates. Biometrika52, 1965, 645–646.
Rao, C.R.: Linear statistical inference and its applications. 2. Aufl. New York-London-Sydney-Toronto 1973.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Stadje, W. Summen von Produkten gemeinsam normalverteilter Zufallsvariablen. Metrika 30, 31–36 (1983). https://doi.org/10.1007/BF02056898
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/BF02056898