Bivariate exponential-type distributions with linear regression |
| |
Authors: | Egmar Rödel |
| |
Institution: | Humboldt-Universit?t , Unter den Linden 6, Berlin, ORZ, DDR-1086 |
| |
Abstract: | Let Xbe a bivariate exponential-type random vector (BIDLIKAR, PATIL (1968)), than it is proved: 1. If P(X ≥0) = 1 is valid, then Xhas linear regression to both directions if and only if Xpossesses a symmetric Γ-distribution. 2. Xpossesses linear regression to both directions with constant regression coefficients (independent of the parameter vector ? of the exponential-type distribution (BIDLIKAR, PATIL (1968)) if and only if Xis normal distributed. |
| |
Keywords: | copula Farlie–Gumbel–Morgenstern distribution sampling procedure |
|
|