Article
Keywords:
band inverse covariance matrix; discriminant analysis; characteristic function; multivariate normal distribution; Wishart distribution; unbiased density estimation; classification rules
Summary:
In a multivariate normal distribution, let the inverse of the covariance matrix be a band matrix. The distribution of the sufficient statistic for the covariance matrix is derived for this case. It is a generalization of the Wishart distribution. The distribution may be used for unbiased density estimation and construction of classification rules.
References:
[1] Abusev R.A., Lumelskiĭ, Ya.P.:
Unbiased estimators and classifìcation problems for multivariate normal populations. Theor. Prob. Appl. 25 (1980), 381-389.
MR 0572573
[3] Ingham A.E.:
An integral which occurs in statistics. Proc. Cambridge Phil. Soc. 29 (1933), 271-276.
Zbl 0007.00701
[4] Lumelskiĭ, Ya.P., Sapozhniкov P.N:
Unbiased estimates of density functions. Theor. Prob. Appl. 14 (1969), 357-365.
DOI 10.1137/1114048
[7] Vapnik V.N.:
Estimation of Dependences Based on Empirical Data. Springer Veгlag, New Yoгk, 1982.
MR 0672244 |
Zbl 0499.62005
[8] Wishart J., Bartlett M.S.:
The generalised product moment distribution in a normal system. Proc. Cambridge Phil. Soc. 29 (1933), 260-270.
Zbl 0007.02304