Merge 850cdb3 into bf1773d

doc/_src_docs/surrogate_models.rstx

@@ -4,22 +4,16 @@ Surrogate modeling methods
 SMT contains the surrogate modeling methods listed below.
 
 .. toctree::
-   :maxdepth: 1
+   :maxdepth: 3
    :titlesonly:
 
+   surrogate_models/ls
+   surrogate_models/qp
+   surrogate_models/gpr
    surrogate_models/rbf
    surrogate_models/idw
    surrogate_models/rmts
-   surrogate_models/ls
-   surrogate_models/qp
-   surrogate_models/krg
-   surrogate_models/kpls
-   surrogate_models/kplsk
-   surrogate_models/gekpls
-   surrogate_models/gpx
    surrogate_models/genn
-   surrogate_models/mgp
-   surrogate_models/sgp
 
 
 Usage

doc/_src_docs/surrogate_models/gpr.rstx

@@ -0,0 +1,23 @@
+Gausian process regression
+==========================
+
+Here below the surrogate models related to Gaussian process regressioh:
+
+* `KRG` implements classic gaussian process regression. 
+* `KPLS` and `KPLSK` are variants using PLS dimension reduction to address high-dimensional training data.
+* `GPX` is a re-implementation of `KRG` and `KPLS` using Rust for faster training/prediction operations.
+* `GEKPLS` leverage derivatives training data to improve the surrogate model quality.
+* `MGP` takes into account the uncertainty of the hyperparameters defined as a density function.   
+* `SGP` implements sparse methods allowing to deal with large training dataset as others implementations have a time complexity of :math:`O(n^3)` as well as a :math:`O(n^2)` memory cost in the number :math:`n` of training points.
+
+.. toctree::
+   :maxdepth: 1
+   :titlesonly:
+
+   gpr/krg
+   gpr/kpls
+   gpr/kplsk
+   gpr/gpx
+   gpr/gekpls
+   gpr/mgp
+   gpr/sgp

doc/_src_docs/surrogate_models/krg.rstx → doc/_src_docs/surrogate_models/gpr/krg.rstx

@@ -1,5 +1,5 @@
-Kriging
-=======
+Kriging (KRG)
+=============
 
 Kriging is an interpolating model that is a linear combination of a known function :math:`f_i({\bf x})` which is added to a realization of a stochastic process :math:`Z({\bf x})`
 

doc/index.rstx

@@ -74,7 +74,7 @@ Documentation contents
 ----------------------
 
 .. toctree::
-   :maxdepth: 2
+   :maxdepth: 3
    :titlesonly:
 
    _src_docs/getting_started

smt/applications/podi.py

@@ -106,7 +106,7 @@ def compute_pod(
             seed number which controls random draws for internal optim. (optional)
 
         Examples
-        ----------
+        --------
         >>> sm.compute_pod(database, tol = 0.99)
         """
         choice_svd = None

smt/surrogate_models/surrogate_model.py

@@ -391,16 +391,16 @@ def predict_variance_derivatives(self, x: np.ndarray, kx: int) -> np.ndarray:
         """
         Provide the derivatives of the variance of the model at a set of points
         Parameters
-        -----------
+        ----------
         x : np.ndarray [n_evals, dim]
             Evaluation point input variable values
         kx : int
             The 0-based index of the input variable with respect to which derivatives are desired.
 
         Returns
         -------
-         derived_variance:  np.ndarray
-             The kx-th derivatives of the variance of the kriging model
+        derived_variance:  np.ndarray
+            The kx-th derivatives of the variance of the kriging model
         """
         check_support(self, "variance_derivatives")
         x = self._pre_predict(x)
@@ -415,14 +415,14 @@ def predict_variance_gradient(self, x: np.ndarray) -> np.ndarray:
         (ie the derivatives wrt to all component at a unique point x)
 
         Parameters
-        -----------
+        ----------
         x : np.ndarray [1, dim] or even (dim,) vector
             Evaluation point input variable values
 
         Returns
         -------
-         derived_variance:  np.ndarray
-             The jacobian of the variance of the kriging model
+        derived_variance :  np.ndarray
+            The jacobian of the variance of the kriging model
         """
         check_support(self, "variance_derivatives")
         if x.shape == (self.nx,):  # allow to pass row vector for convenience