Index: trunk/SDToolBox/statistical_model.py
===================================================================
diff -u -r100 -r101
--- trunk/SDToolBox/statistical_model.py	(.../statistical_model.py)	(revision 100)
+++ trunk/SDToolBox/statistical_model.py	(.../statistical_model.py)	(revision 101)
@@ -14,6 +14,10 @@
 from sklearn.model_selection import train_test_split
 from sklearn.model_selection import TimeSeriesSplit
 from sklearn.linear_model import LinearRegression
+from sklearn.metrics import explained_variance_score
+from sklearn.metrics import mean_squared_error
+from scipy.stats import pearsonr
+from scipy.stats import ks_2samp
 from sklearn import datasets
 from sklearn import preprocessing
 from sklearn import svm
@@ -120,29 +124,43 @@
             scaler.transform(test_model)
 
     @staticmethod
-    def robutness_metrics(train_model, test_model) -> Tuple[float, float]:
+    def robutness_metrics(obs: np.array, sim: np.array) -> Dict[str, float]:
         """Compares terms of bias, ratio variances, RMSE,
         correlation coefficient and validity of the distribution
         using Kolmogorov-Smirnov
 
         Arguments:
-            train_model {np.Array} -- Trained model.
-            test_model {np.Array} -- Test / validation model.
+            obs {np.Array} -- Original data split validation.
+            sim {np.Array} -- Regression result model.
 
         Returns:
-            Tuple[float,float] -- (D, p-value)
+            Dict[str, float] -- Dictionary of metrics and their results.
         """
-        return stats.ks_2samp(train_model, test_model)
+        return {
+            'BIAS': StatisticalModel.get_BIAS(obs, sim, 0),
+            'RMSE': np.sqrt(mean_squared_error(obs, sim, multioutput='raw_values')),
+            'Ratio variances': obs.std()/sim.std(),
+            'Explained variance': explained_variance_score(obs, sim, multioutput='raw_values'),
+            'Correlation coefficient': pearsonr(obs, sim),
+            'Distribution validation': ks_2samp(obs, sim),
+        }
 
     @staticmethod
-    def multivariate_mlr(X, Y):
-        """Multivariate multiple linear regression.
+    def get_BIAS(
+            measured_sample: np.array,
+            predicted_sample: np.array,
+            ax: int) -> np.array:
+        """Calculates the bias between two samples.
 
         Arguments:
-            X {[np.Array]} -- Reference predictor data.
+            measured_sample {np.array} -- Measured sample.
+            predicted_sample {np.array} -- Predicted sample.
+            ax {int} -- Axis for the n_samples.
+
+        Returns:
+            np.array -- Calculated bias based on the input samples.
         """
-        # Y = sum(alpha_i + X_i)
-        raise Exception(om.error_function_not_implemented)
+        return np.mean(predicted_sample, axis=ax) - np.mean(measured_sample, axis=ax)
 
     @staticmethod
     def bias_and_variance_correction(train_model):
@@ -157,3 +175,57 @@
         scaler = preprocessing.StandardScaler().fit(train_model)
         projection = scaler.transform(train_model)
         return (projection - projection.mean(axis=0))/(projection.std(axis=0))*train_model.std(axis=0) + train_model.mean(axis=0)
+
+    @staticmethod
+    def multivariate_mlr(X: np.array, Y: np.array):
+        """Multivariate multiple linear regression.
+        """
+        if(len(X) != len(Y)):
+            raise Exception('Argument arrays should have the same size.')
+
+        alpha_coeff = [Y[i] - X[i] for i in range(0, len(X))]
+
+        raise Exception(om.error_function_not_implemented)
+
+
+    @staticmethod
+    def optimum_regression_model_ess(X: np.array, Y: np.array):
+        raise Exception(om.error_function_not_implemented)
+
+    @staticmethod
+    def __get_angular_components(component: np.array) -> Tuple[np.array, np.array]:
+        """Gets the angular components (cos / sin) for the given component.
+
+        Arguments:
+            component {np.array} -- Component to calculate.
+
+        Returns:
+            Tuple[np.array, np.array] -- Calculated angular components.
+        """
+        pi = math.pi
+        temp = pi/2 - component*pi/180
+        neg_idx = [idx for idx in range(0, len(temp)) if temp[idx] < -pi]
+        # Correct indices.
+        for idx in neg_idx:
+            temp[idx] = 2*pi+temp[idx]
+
+        return math.cos(temp), math.sin(temp)
+
+    @staticmethod
+    def regression_model_circular(X: np.array, Y: np.array):
+        raise Exception(om.error_function_not_implemented)
+        # preinitialize
+        # Ycalx, Ycaly = StatisticalModel.__get_angular_components(X)
+
+        # # Regression
+        # Xcal
+
+        # # Validation
+
+        # # Linear regression
+        # regression = LinearRegression().fit(X, Y)
+        # b = regress(Ycal, Xcal)
+        # Ymod = [np.ones(len(Xcal))] * b[0]
+        # Ymod = [ Ymod[i] + b[i+1]*Xcal[:, i+1] for i in range(0, len(b)-1)]
+        # result = Ycal - Ymod
+        # return sum(np.square(result)), result