Stan Math Library
5.0.0
Automatic Differentiation
|
return_type_t< T_y, T_dof, T_scale > stan::math::inv_wishart_lpdf | ( | const T_y & | W, |
const T_dof & | nu, | ||
const T_scale & | S | ||
) |
The log of the Inverse-Wishart density for the given W, degrees of freedom, and scale matrix.
The scale matrix, S, must be k x k, symmetric, and semi-positive definite.
\begin{eqnarray*} W &\sim& \mbox{\sf{Inv-Wishart}}_{\nu} (S) \\ \log (p (W \, |\, \nu, S) ) &=& \log \left( \left(2^{\nu k/2} \pi^{k (k-1) /4} \prod_{i=1}^k{\Gamma (\frac{\nu + 1 - i}{2})} \right)^{-1} \times \left| S \right|^{\nu/2} \left| W \right|^{-(\nu + k + 1) / 2} \times \exp (-\frac{1}{2} \mbox{tr} (S W^{-1})) \right) \\ &=& -\frac{\nu k}{2}\log(2) - \frac{k (k-1)}{4} \log(\pi) - \sum_{i=1}^{k}{\log (\Gamma (\frac{\nu+1-i}{2}))} +\frac{\nu}{2} \log(\det(S)) - \frac{\nu+k+1}{2}\log (\det(W)) - \frac{1}{2} \mbox{tr}(S W^{-1}) \end{eqnarray*}
T_y | type of scalar |
T_dof | type of degrees of freedom |
T_scale | type of scale |
W | A scalar matrix |
nu | Degrees of freedom |
S | The scale matrix |
std::domain_error | if nu is not greater than k-1 |
std::domain_error | if S is not square, not symmetric, or not semi-positive definite. |
Definition at line 45 of file inv_wishart_lpdf.hpp.