Implement OneNearestNeighbor classifier

numpy_questions.py

@@ -19,49 +19,50 @@
 
 
 def max_index(X):
-    """Return the index of the maximum in a numpy array.
+    """Return the index of the maximum in a 2D numpy array.
 
     Parameters
     ----------
-    X : ndarray of shape (n_samples, n_features)
-        The input array.
+    X : ndarray of shape (n_rows, n_cols)
+        The input 2D array.
 
     Returns
     -------
     (i, j) : tuple(int)
-        The row and columnd index of the maximum.
+        The row and column index of the maximum value in X.
 
     Raises
     ------
     ValueError
-        If the input is not a numpy array or
-        if the shape is not 2D.
+        If the input is not a numpy array or is not 2D.
     """
-    i = 0
-    j = 0
+    if not isinstance(X, np.ndarray):
+        raise ValueError("Input must be a numpy array.")
+    if X.ndim != 2:
+        raise ValueError("Input array must be 2-dimensional.")
 
-    # TODO
-
-    return i, j
+    flat_index = np.argmax(X)
+    return np.unravel_index(flat_index, X.shape)
 
 
 def wallis_product(n_terms):
     """Implement the Wallis product to compute an approximation of pi.
 
-    See:
-    https://en.wikipedia.org/wiki/Wallis_product
-
     Parameters
     ----------
     n_terms : int
-        Number of steps in the Wallis product. Note that `n_terms=0` will
-        consider the product to be `1`.
+        Number of steps in the Wallis product. When n_terms = 0, the
+        product is defined as 1.0.
 
     Returns
     -------
     pi : float
-        The approximation of order `n_terms` of pi using the Wallis product.
+        Approximation of pi using the Wallis product.
     """
-    # XXX : The n_terms is an int that corresponds to the number of
-    # terms in the product. For example 10000.
-    return 0.
+    if n_terms == 0:
+        return 1.0
+
+    k = np.arange(1, n_terms + 1, dtype=float)
+    factors = (4 * k**2) / (4 * k**2 - 1)
+    product = np.prod(factors)
+    return float(2 * product)

sklearn_questions.py

@@ -29,46 +29,97 @@
 
 
 class OneNearestNeighbor(BaseEstimator, ClassifierMixin):
-    "OneNearestNeighbor classifier."
+    """One-nearest-neighbor classifier.
+
+    This classifier predicts the label of a sample using the label of the
+    closest training sample according to the Euclidean distance.
+    """
 
     def __init__(self):  # noqa: D107
+        # Pas d'hyperparamètres pour ce modèle
         pass
 
     def fit(self, X, y):
-        """Write docstring.
-
-        And describe parameters
+        """Fit the OneNearestNeighbor classifier.
+
+        Parameters
+        ----------
+        X : array-like of shape (n_samples, n_features)
+            Training data.
+        y : array-like of shape (n_samples,)
+            Target labels.
+
+        Returns
+        -------
+        self : OneNearestNeighbor
+            Fitted estimator.
         """
+        # Vérifications standard scikit-learn
         X, y = check_X_y(X, y)
         check_classification_targets(y)
+
+        # Attributs nécessaires pour scikit-learn
         self.classes_ = np.unique(y)
         self.n_features_in_ = X.shape[1]
 
-        # XXX fix
+        # On mémorise les données d'entraînement
+        self.X_ = X
+        self.y_ = y
+
         return self
 
     def predict(self, X):
-        """Write docstring.
+        """Predict class labels for the given test data.
+
+        Parameters
+        ----------
+        X : array-like of shape (n_samples, n_features)
+            Test samples.
 
-        And describe parameters
+        Returns
+        -------
+        y_pred : ndarray of shape (n_samples,)
+            Predicted class labels for each sample in X.
         """
         check_is_fitted(self)
         X = check_array(X)
-        y_pred = np.full(
-            shape=len(X), fill_value=self.classes_[0],
-            dtype=self.classes_.dtype
-        )
 
-        # XXX fix
+        # Vérifier que le nombre de features correspond
+        if X.shape[1] != self.n_features_in_:
+            raise ValueError("Number of features of X does not match training")
+
+        n_samples_test = X.shape[0]
+        y_pred = np.empty(n_samples_test, dtype=self.y_.dtype)
+
+        # Pour chaque point de test, on cherche le point d'entraînement
+        # le plus proche (distance euclidienne) et on copie son label.
+        for i in range(n_samples_test):
+            # différences entre x_i et tous les X_ de train
+            diffs = self.X_ - X[i]
+            # distances euclidiennes (norme L2)
+            dists = np.linalg.norm(diffs, axis=1)
+            # index du plus proche voisin
+            nearest_index = np.argmin(dists)
+            y_pred[i] = self.y_[nearest_index]
+
         return y_pred
 
     def score(self, X, y):
-        """Write docstring.
-
-        And describe parameters
+        """Return the mean accuracy on the given test data and labels.
+
+        Parameters
+        ----------
+        X : array-like of shape (n_samples, n_features)
+            Test samples.
+        y : array-like of shape (n_samples,)
+            True labels for X.
+
+        Returns
+        -------
+        score : float
+            Mean accuracy of predictions on X compared to y.
         """
         X, y = check_X_y(X, y)
         y_pred = self.predict(X)
-
-        # XXX fix
-        return y_pred.sum()
+        # proportion de bonnes prédictions
+        return np.mean(y_pred == y)