calibration not improves

src/author_identification.py

@@ -57,7 +57,7 @@ def main():
 
         print('Fitting the Verificator')
         if args.C is None:
-            params = {'C': np.logspace(-4, +3, 8)}
+            params = {'C': np.logspace(-3, +3, 7)}
             C = 1.
         else:
             params = None

src/model.py

@@ -36,8 +36,8 @@ class AuthorshipVerificator:
             print(f'Best params: {self.estimator.best_params_} (cross-validation F1={f1_mean:.3f})')
             self.estimator = self.estimator.best_estimator_
 
-        self.estimator = CalibratedClassifierCV(base_estimator=self.estimator, cv=self.nfolds, ensemble=False)
-        self.estimator.fit(X, y)
+        #self.estimator = CalibratedClassifierCV(base_estimator=self.estimator, cv=self.nfolds, ensemble=False)
+        #self.estimator.fit(X, y)
 
         return self
 

src/util/calibration.py

@@ -23,7 +23,7 @@ from sklearn.base import (BaseEstimator, ClassifierMixin, RegressorMixin, clone,
 from sklearn.preprocessing import label_binarize, LabelBinarizer
 from sklearn.utils import check_array, indexable, column_or_1d
 from sklearn.utils.validation import check_is_fitted, check_consistent_length
-from sklearn.utils.validation import _check_sample_weight
+#from sklearn.utils.validation import _check_sample_weight
 from sklearn.isotonic import IsotonicRegression
 from sklearn.svm import LinearSVC
 from sklearn.model_selection import check_cv, cross_val_predict
@@ -586,4 +586,4 @@ def calibration_curve(y_true, y_prob, *, normalize=False, n_bins=5,
     prob_true = bin_true[nonzero] / bin_total[nonzero]
     prob_pred = bin_sums[nonzero] / bin_total[nonzero]
 
-    return prob_true, prob_pred
+    return prob_true, prob_pred