1 files changed, 76 insertions, 0 deletions
diff --git a/inst/model/prior_struct/stk_predict_.m b/inst/model/prior_struct/stk_predict_.m
new file mode 100644
index 0000000..8f05133
--- /dev/null
+++ b/inst/model/prior_struct/stk_predict_.m
@@ -0,0 +1,76 @@
+% STK_PREDICT_ [STK internal]
+%
+% See also: stk_predict
+
+% Copyright Notice
+%
+%    Copyright (C) 2020 CentraleSupelec
+%
+%    Author:  Julien Bect  <julien.bect@centralesupelec.fr>
+
+% Copying Permission Statement
+%
+%    This file is part of
+%
+%            STK: a Small (Matlab/Octave) Toolbox for Kriging
+%               (https://github.com/stk-kriging/stk/)
+%
+%    STK is free software: you can redistribute it and/or modify it under
+%    the terms of the GNU General Public License as published by the Free
+%    Software Foundation,  either version 3  of the License, or  (at your
+%    option) any later version.
+%
+%    STK is distributed  in the hope that it will  be useful, but WITHOUT
+%    ANY WARRANTY;  without even the implied  warranty of MERCHANTABILITY
+%    or FITNESS  FOR A  PARTICULAR PURPOSE.  See  the GNU  General Public
+%    License for more details.
+%
+%    You should  have received a copy  of the GNU  General Public License
+%    along with STK.  If not, see <http://www.gnu.org/licenses/>.
+
+function [zp, lambda, mu, K] = stk_predict_ (model, xt)
+
+% If we end up here, model must be a prior model struct (as opposed to a
+% model object).  Better safe than sorry:
+stk_assert_model_struct (model);
+
+nt = stk_get_sample_size (xt);
+
+% Is it a proper or an improper prior?
+model = stk_model_fixlm (model);
+if isa (model.lm, 'stk_lm_null')
+    % Proper prior (currently, zero-mean only)
+    zp_mean = zeros (nt, 1);
+else
+    % Improper prior
+    zp_mean = nan (nt, 1);
+end
+
+if nargout > 3
+    K = stk_make_matcov (model, xt, xt, false);
+    zp_var = diag (K);
+else
+    zp_var = stk_make_matcov (model, xt, xt, true);
+end
+
+zp = stk_dataframe ([zp_mean zp_var], {'mean' 'var'});
+
+if nargout > 1  % We want lambda
+
+    % lambda must be n x nt, with n the sample size (here n = 0)
+    lambda = zeros (0, nt);
+    
+    if nargout > 2  % We want mu as well
+        
+        % UGLY: we have no cleaner way to find the dimension...
+        x_ = xt(1, :);
+        [K_ignore, P] = stk_make_matcov (model, x_, x_);  %#ok<ASGLU>
+        r = size (P, 2);
+        
+        % mu must be r x nt, and setting it to zero in this case makes
+        % sense because the posterior variance formula remains valid
+        mu = zeros (r, nt);
+    end
+end
+
+end % function