GP.logL_hessianable

GP.logL_hessianable(p: Any, Y: Array) → Any

Computes the log likelihood without returning any stored values from the decomposition of the covariance matrix. This function is slower for gradient calculations than GP.logL but is more numerically stable for second-order derivative calculations as required when calculating the hessian. This function still only returns the log likelihood so jax.hessian must be applied to return the hessian of the log likelihood.

Parameters:

• p (PyTree) – Pytree of hyperparameters used to calculate the covariance matrix in addition to any mean function parameters which may be needed to calculate the mean function.

• Y (JAXArray) – Observed data to fit, must be of shape (N_l, N_t).

Returns:

The value of the log likelihood.

Return type:

Scalar