doc: gp_inf

Bayesian system resolution for computing posterior of GP given the observations Yt at Xt


 BayesInv = gp_inf(Ktt, Yt, noise)
 BayesInv = gp_inf(Ktt, Yt, noise, Ht)


  • Ktt matrix (nt, nt) of kernel between the points of Xt
  • Yt vector (nt, 1) of observations
  • noise noise standard deviation $\eta^2$
  • Ht matrix (nt, b) of basis data as returned by basis_cst(Xt)


struct array containing:

  • RC upper triangular matrix (nt,nt) of Cholesky decomposition of Ktt+noise*I
  • invCY vector (nt,1) solution of $(K + \eta^2 I)^{-1} Y$
  • beta vector (b,1) solution of the basis system

See also

gp_pred | gp_inf_update