gp_exponential_cov() [6/6]

template<typename T_x1 , typename T_x2 , typename T_s , typename T_l >

Eigen::Matrix< return_type_t< T_x1, T_x2, T_s, T_l >, Eigen::Dynamic, Eigen::Dynamic > stan::math::gp_exponential_cov ( const std::vector< Eigen::Matrix< T_x1, -1, 1 > > &  x1, const std::vector< Eigen::Matrix< T_x2, -1, 1 > > &  x2, const T_s &  sigma, const std::vector< T_l > &  length_scale  )

inline

Returns a Matern exponential cross covariance matrix.

\[ k(x, x') = \sigma^2 exp(-\sum_{k=1}^K\frac{d(x, x')}{l_k}) \]

where \(d(x, x')\) is the Euclidean distance

This function is for the cross covariance matrix needed to compute the posterior predictive density.

Template Parameters

T_x1	first type of std::vector of scalars
T_x2	second type of std::vector of scalars
T_s	type of parameter sigma, marginal standard deviation
T_l	type of parameter length scale

Parameters

x1	std::vector of Eigen vectors of scalars
x2	std::vector of Eigen vectors of scalars
length_scale	parameter length scale
sigma	standard deviation that can be used in stan::math::square

Exceptions

std::domain

error if sigma <= 0, l <= 0, or x1, x2 are nan or inf

Definition at line 248 of file gp_exponential_cov.hpp.