mlquantify 0.1.13__tar.gz → 0.1.15__tar.gz

{mlquantify-0.1.13/mlquantify.egg-info → mlquantify-0.1.15}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mlquantify
-Version: 0.1.13
+Version: 0.1.15
 Summary: Quantification Library
 Home-page: https://github.com/luizfernandolj/QuantifyML/tree/master
 Maintainer: Luiz Fernando Luth Junior

mlquantify-0.1.15/VERSION.txt

@@ -0,0 +1 @@
+0.1.15

{mlquantify-0.1.13 → mlquantify-0.1.15}/mlquantify/adjust_counting/_adjustment.py

@@ -2,6 +2,7 @@ import numpy as np
 from abc import abstractmethod
 from scipy.optimize import minimize
 import warnings
+from sklearn.metrics import confusion_matrix
 
 from mlquantify.adjust_counting._base import BaseAdjustCount
 from mlquantify.adjust_counting._counting import CC, PCC
@@ -104,7 +105,7 @@ class ThresholdAdjustment(SoftLearnerQMixin, BaseAdjustCount):
         thresholds, tprs, fprs = evaluate_thresholds(train_y_values, positive_scores)
         threshold, tpr, fpr = self.get_best_threshold(thresholds, tprs, fprs)
 
-        cc_predictions = CC(threshold).aggregate(predictions)[1]
+        cc_predictions = CC(threshold=threshold).aggregate(predictions, train_y_values)[1]
 
         if tpr - fpr == 0:
             prevalence = cc_predictions
@@ -208,7 +209,7 @@ class MatrixAdjustment(BaseAdjustCount):
             prevs_estim = self._get_estimations(predictions > priors)
             prevalence = self._solve_optimization(prevs_estim, priors)
         else:
-            self.CM = self._compute_confusion_matrix(train_y_pred)
+            self.CM = self._compute_confusion_matrix(train_y_pred, train_y_values)
             prevs_estim = self._get_estimations(predictions)
             prevalence = self._solve_linear(prevs_estim)
         
@@ -389,8 +390,11 @@ class GAC(CrispLearnerQMixin, MatrixAdjustment):
     def __init__(self, learner=None):
         super().__init__(learner=learner, solver='linear')
     
-    def _compute_confusion_matrix(self, predictions):
-        prev_estim = self._get_estimations(predictions)
+    def _compute_confusion_matrix(self, predictions, y_values):
+        self.CM = confusion_matrix(y_values, predictions, labels=self.classes_).T
+        self.CM = self.CM.astype(float)
+        prev_estim = self.CM.sum(axis=0)
+
         for i, _ in enumerate(self.classes_):
             if prev_estim[i] == 0:
                 self.CM[i, i] = 1
@@ -448,13 +452,16 @@ class GPAC(SoftLearnerQMixin, MatrixAdjustment):
     def __init__(self, learner=None):
         super().__init__(learner=learner, solver='linear')
     
-    def _compute_confusion_matrix(self, posteriors):
-        prev_estim = self._get_estimations(posteriors)
-        for i, _ in enumerate(self.classes_):
-            if prev_estim[i] == 0:
-                self.CM[i, i] = 1
-            else:
-                self.CM[:, i] /= prev_estim[i]
+    def _compute_confusion_matrix(self, posteriors, y_values):
+        n_classes = len(self.classes_)
+        confusion = np.eye(n_classes)
+
+        for i, class_label in enumerate(self.classes_):
+            indices = (y_values == class_label)
+            if np.any(indices):
+                confusion[i] = posteriors[indices].mean(axis=0)
+        
+        self.CM = confusion.T
         return self.CM
 
 
@@ -514,7 +521,7 @@ class X_method(ThresholdAdjustment):
            *ECML*, pp. 564-575.
     """
     def get_best_threshold(self, thresholds, tprs, fprs):
-        idx = np.argmin(np.abs(1 - (tprs + fprs)))
+        idx = np.argmin(np.abs((1-tprs) - fprs))
         return thresholds[idx], tprs[idx], fprs[idx]
 
 
@@ -601,7 +608,7 @@ class MS(ThresholdAdjustment):
         
         prevs = []
         for thr, tpr, fpr in zip(thresholds, tprs, fprs):
-            cc_predictions = CC(thr).aggregate(predictions)
+            cc_predictions = CC(threshold=thr).aggregate(predictions, train_y_values)
             cc_predictions = cc_predictions[1]
             
             if tpr - fpr == 0:

{mlquantify-0.1.13 → mlquantify-0.1.15}/mlquantify/adjust_counting/_base.py

@@ -204,7 +204,7 @@ class BaseAdjustCount(AggregationMixin, BaseQuantifier):
         self.learner = learner
 
     @_fit_context(prefer_skip_nested_validation=True)
-    def fit(self, X, y, learner_fitted=False):
+    def fit(self, X, y, learner_fitted=False, cv=10, stratified=True, random_state=None, shuffle=True):
         """Fit the quantifier using the provided data and learner."""
         X, y = validate_data(self, X, y)
         validate_y(self, y)
@@ -220,10 +220,10 @@ class BaseAdjustCount(AggregationMixin, BaseQuantifier):
                 X,
                 y,
                 function=learner_function,
-                cv=5,
-                stratified=True,
-                random_state=None,
-                shuffle=True
+                cv=cv,
+                stratified=stratified,
+                random_state=random_state,
+                shuffle=shuffle
             )
         
         self.train_predictions = train_predictions
@@ -241,7 +241,6 @@ class BaseAdjustCount(AggregationMixin, BaseQuantifier):
         self.classes_ = check_classes_attribute(self, np.unique(y_train_values))
         
         predictions = validate_predictions(self, predictions)
-        train_predictions = validate_predictions(self, train_predictions)
         
         prevalences = self._adjust(predictions, train_predictions, y_train_values)
         prevalences = validate_prevalences(self, prevalences, self.classes_)

{mlquantify-0.1.13 → mlquantify-0.1.15}/mlquantify/adjust_counting/_counting.py

@@ -75,10 +75,12 @@ class CC(CrispLearnerQMixin, BaseCount):
         super().__init__(learner=learner)
         self.threshold = threshold
 
-    def aggregate(self, predictions):
-        predictions = validate_predictions(self, predictions)
+    def aggregate(self, predictions, train_y_values=None):
+        predictions = validate_predictions(self, predictions, self.threshold)
         
-        self.classes_ = check_classes_attribute(self, np.unique(predictions))
+        if train_y_values is None:
+            train_y_values = np.unique(predictions)
+        self.classes_ = check_classes_attribute(self, np.unique(train_y_values))
         class_counts = np.array([np.count_nonzero(predictions == _class) for _class in self.classes_])
         prevalences = class_counts / len(predictions)
         

{mlquantify-0.1.13 → mlquantify-0.1.15}/mlquantify/meta/_classes.py

@@ -776,9 +776,9 @@ class QuaDapt(MetaquantifierMixin, BaseQuantifier):
         n_neg = n - n_pos
         
         # Scores positivos
-        p_score = np.random.uniform(size=n_pos) ** merging_factor
+        p_score = np.random.uniform(size=n_pos) ** m
         # Scores negativos
-        n_score = 1 - (np.random.uniform(size=n_neg) ** merging_factor)
+        n_score = 1 - (np.random.uniform(size=n_neg) ** m)
         
         # Construção dos arrays de features (duas colunas iguais)
         moss = np.column_stack(

{mlquantify-0.1.13 → mlquantify-0.1.15}/mlquantify/utils/_validation.py

@@ -97,22 +97,22 @@ def validate_y(quantifier: Any, y: np.ndarray) -> None:
 def _get_valid_crisp_predictions(predictions, threshold=0.5):
     predictions = np.asarray(predictions)
 
-    dimensions = predictions.shape[1] if len(predictions.shape) > 1 else 1
+    dimensions = predictions.ndim
 
     if dimensions > 2:
         predictions = np.argmax(predictions, axis=1)
     elif dimensions == 2:
-        predictions = (predictions[:, 1] > threshold).astype(int)
+        predictions = (predictions[:, 1] >= threshold).astype(int)
     elif dimensions == 1:
         if np.issubdtype(predictions.dtype, np.floating):
-            predictions = (predictions > threshold).astype(int)
+            predictions = (predictions >= threshold).astype(int)
     else:
         raise ValueError(f"Predictions array has an invalid number of dimensions. Expected 1 or more dimensions, got {predictions.ndim}.")
 
     return predictions
 
 
-def validate_predictions(quantifier: Any, predictions: np.ndarray) -> None:
+def validate_predictions(quantifier: Any, predictions: np.ndarray, threshold: float = 0.5) -> np.ndarray:
     """
     Validate predictions using the quantifier's declared output tags.
     Raises InputValidationError if inconsistent with tags.
@@ -132,7 +132,7 @@ def validate_predictions(quantifier: Any, predictions: np.ndarray) -> None:
             f"Soft predictions for {quantifier.__class__.__name__} must be float, got dtype {predictions.dtype}."
         )
     elif estimator_type == "crisp" and np.issubdtype(predictions.dtype, np.floating):
-        predictions = _get_valid_crisp_predictions(predictions) 
+        predictions = _get_valid_crisp_predictions(predictions, threshold) 
     return predictions   
     
     

{mlquantify-0.1.13 → mlquantify-0.1.15/mlquantify.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mlquantify
-Version: 0.1.13
+Version: 0.1.15
 Summary: Quantification Library
 Home-page: https://github.com/luizfernandolj/QuantifyML/tree/master
 Maintainer: Luiz Fernando Luth Junior

mlquantify-0.1.13/VERSION.txt

@@ -1 +0,0 @@
-0.1.13