1. """
2. ================================
3. Nearest Neighbors Classification
4. ================================
5.  
6. Sample usage of Nearest Neighbors classification.
7. It will plot the decision boundaries for each class.
8. """
9. print(__doc__)
10.  
11. import numpy as np
12. import matplotlib.pyplot as plt
13. from matplotlib.colors import ListedColormap
14. from sklearn import neighbors, datasets
15.  
16. n_neighbors = 15
17.  
18. # import some data to play with
19. iris = datasets.load_iris()
20.  
21. # we only take the first two features. We could avoid this ugly
22. # slicing by using a two-dim dataset
23. X = iris.data[:, :2]
24. y = iris.target
25.  
26. h = .02  # step size in the mesh
27.  
28. # Create color maps
29. cmap_light = ListedColormap(['#FFAAAA', '#AAFFAA', '#AAAAFF'])
30. cmap_bold = ListedColormap(['#FF0000', '#00FF00', '#0000FF'])
31.  
32. for weights in ['uniform', 'distance']:
33.     # we create an instance of Neighbours Classifier and fit the data.
34.     clf = neighbors.KNeighborsClassifier(n_neighbors, weights=weights)
35.     clf.fit(X, y)
36.  
37.     # Plot the decision boundary. For that, we will assign a color to each
38.     # point in the mesh [x_min, x_max]x[y_min, y_max].
39.     x_min, x_max = X[:, 0].min() - 1, X[:, 0].max() + 1
40.     y_min, y_max = X[:, 1].min() - 1, X[:, 1].max() + 1
41.     xx, yy = np.meshgrid(np.arange(x_min, x_max, h),
42.                          np.arange(y_min, y_max, h))
43.     Z = clf.predict(np.c_[xx.ravel(), yy.ravel()])
44.  
45.     # Put the result into a color plot
46.     Z = Z.reshape(xx.shape)
47.     plt.figure()
48.     plt.pcolormesh(xx, yy, Z, cmap=cmap_light)
49.  
50.     # Plot also the training points
51.     plt.scatter(X[:, 0], X[:, 1], c=y, cmap=cmap_bold,
52.                 edgecolor='k', s=20)
53.     plt.xlim(xx.min(), xx.max())
54.     plt.ylim(yy.min(), yy.max())
55.     plt.title("3-Class classification (k = %i, weights = '%s')"
56.               % (n_neighbors, weights))
57.  
58. plt.show()