Greatest Lower Bound to the Elliptical Theory Kurtosis Parameter
证明了连续p维椭圆分布（协方差矩阵为正定实矩阵）的峰度参数K必须严格大于-2/(p+2)，为统计建模提供了理论下界。
It is shown that the kurtosis parameter K for a continuous p-variate elliptical distribution with real positive-definite covariance matrix Ω must be strictly greater than −2/(p+2).