mkl

multimodal/datasets/digit.py

+from sklearn import datasets
+import numpy as np
+import PIL
+import matplotlib.pyplot as plt
+import os
+import matplotlib.pyplot as plt
+from multimodal.datasets.base import load_dict, save_dict
+from multimodal.tests.data.get_dataset_path import get_dataset_path
+from multimodal.datasets.data_sample import MultiModalArray
+from multimodal.kernels.mvml import MVML
+#Load the digits dataset
+digits = datasets.load_digits()
+
+#Display the first digit
+plt.figure(1, figsize=(3, 3))
+plt.imshow(digits.images[-1], cmap=plt.cm.gray_r, interpolation='nearest')
+plt.show()
+colors = digits.data
+gradiant = np.gradient(digits.images, axis=[1,2])
+print(gradiant[0].shape)
+gradiant0 = gradiant[0].reshape(colors.shape[0], colors.shape[1])
+gradiant1 = gradiant[1].reshape(colors.shape[0], colors.shape[1])
+for ind in range(digits.images.shape[0]):
+    ima0 = digits.images[ind, :,:]
+    ima1 = gradiant[0][ind, :,:]
+    ima2 = gradiant[1][ind, :,:]
+    ama_pil0 = PIL.Image.fromarray(ima0, mode=None)
+    ama_pil1 = PIL.Image.fromarray(ima1, mode=None)
+    ama_pil2 = PIL.Image.fromarray(ima2, mode=None)
+    histo_color = np.asarray(ama_pil0.histogram())
+    histo_gradiant0 = np.asarray(ama_pil1.histogram())
+    histo_gradiant1 = np.asarray(ama_pil2.histogram())
+    if ind==0:
+        list_histogram_color = histo_color
+        list_histogram_gradiant0 = histo_gradiant0
+        list_histogram_gradiant1 = histo_gradiant1
+    else:
+        list_histogram_color = np.vstack((list_histogram_color, histo_color))
+        list_histogram_gradiant0 = np.vstack((list_histogram_gradiant0, histo_gradiant0))
+        list_histogram_gradiant1 = np.vstack((list_histogram_gradiant1, histo_gradiant1))
+
+dict_digit = {0: list_histogram_color, 1: list_histogram_gradiant0, 2: list_histogram_gradiant1}
+
+
+print(list_histogram_color.shape)
+print(list_histogram_gradiant0.shape)
+print(list_histogram_gradiant1.shape)
+file = get_dataset_path("digit_histogram.npy")
+save_dict(dict_digit, file)
+
+d2 = load_dict(file)
+
+figure = plt.figure(figsize=(27, 9))
+ax = plt.subplot(2,1,1)
+
+ax.scatter(list_histogram_color[:,3], list_histogram_color[:,4], c=digits.target, edgecolors='k')
+ax = plt.subplot(2,1,2)
+ax.scatter(list_histogram_color[:,0], list_histogram_color[:,1], c=digits.target, edgecolors='k')
+plt.show()
+
+mvml = MVML(lmbda=0.1, eta=1, nystrom_param=0.2)
+mvml.fit(d2)

multimodal/kernels/lpMKL.py

@@ -6,6 +6,8 @@ from sklearn.utils.validation import check_X_y
 from sklearn.utils.validation  import check_array
 from sklearn.utils.validation  import check_is_fitted
 from multimodal.kernels.mkernel import MKernel
+from sklearn.utils.multiclass import type_of_target
+from sklearn.utils.multiclass import check_classification_targets
 
 
 class MKL(BaseEstimator, ClassifierMixin, MKernel):
@@ -122,9 +124,18 @@ class MKL(BaseEstimator, ClassifierMixin, MKernel):
         self : object
             Returns self.
         """
-        self.X_, self.K_ = self._global_kernel_transform(X, views_ind=views_ind)
-        self.classes_ = unique_labels(y)
+        self.X_, self.K_= self._global_kernel_transform(X, views_ind=views_ind)
         check_X_y(self.X_, y)
+        check_classification_targets(y)
+        if type_of_target(y) in "binary":
+            self.classes_, y = np.unique(y, return_inverse=True)
+            y[y==0] = -1.0
+        elif type_of_target(y) in "continuous":
+            y = y.astype(float)
+            self.regression_ = True
+        else:
+            raise ValueError("MKL algorithms is a binary classifier"
+                             " or performs regression with float target")
         self.y_ = y
         n = self.K_.shape[0]
         self._calc_nystrom(self.K_, n)
@@ -270,13 +281,18 @@ class MKL(BaseEstimator, ClassifierMixin, MKernel):
             Predicted classes.
         """
         check_is_fitted(self, ['X_', 'C', 'K_', 'y_', 'weights'])
-        X , test_kernels = self._global_kernel_transform(X,
+        X, K = self._global_kernel_transform(X,
                                                          views_ind=self.X_.views_ind,
                                                          Y=self.X_)
         check_array(X)
         C = self.C
         weights  = self.weights
-        return self.lpMKL_predict(test_kernels, C, weights)
+        pred = self.lpMKL_predict(K, C, weights)
+
+        pred = np.sign(pred)
+        pred = pred.astype(int)
+        pred = np.where(pred == -1, 0, pred)
+        return np.take(self.classes_, pred)
 
 
     def lpMKL_predict(self, X, C, weights):

multimodal/kernels/mvml.py

@@ -4,6 +4,7 @@ from scipy.sparse.linalg import splu
 from scipy.sparse import csc_matrix
 from sklearn.base import BaseEstimator
 from sklearn.base import ClassifierMixin
+from sklearn.base import RegressorMixin
 from sklearn.utils.multiclass import unique_labels
 from sklearn.metrics.pairwise import pairwise_kernels
 from sklearn.utils.validation import check_X_y
@@ -25,7 +26,7 @@ in International Conference on Artificial Intelligence and Statistics (AISTATS)
 """
 
 
-class MVML(MKernel, BaseEstimator, ClassifierMixin):
+class MVML(MKernel, BaseEstimator, ClassifierMixin, RegressorMixin):
     r"""
     The MVML Classifier
 
@@ -426,16 +427,7 @@ class MVML(MKernel, BaseEstimator, ClassifierMixin):
               for each view.
             - {array like} with (n_samples, nviews *  n_features) with 'views_ind' diferent to 'None'
 
-        views_ind : array-like (default=[0, n_features//2, n_features])
-            Paramater specifying how to extract the data views from X:
 
-            - views_ind is a 1-D array of sorted integers, the entries
-              indicate the limits of the slices used to extract the views,
-              where view ``n`` is given by
-              ``X[:, views_ind[n]:views_ind[n+1]]``.
-
-              With this convention each view is therefore a view (in the NumPy
-              sense) of X and no copy of the data is done.
 
         Returns
         -------