SVM Margins Example

The plots below illustrate the effect the parameter C has on the separation line. A large value of C basically tells our model that we do not have that much faith in our data?s distribution, and will only consider points close to line of separation.

A small value of C includes more/all the observations, allowing the margins to be calculated using all the data in the area.

print(__doc__)
 
 
# Code source: Ga Varoquaux
# Modified for documentation by Jaques Grobler
# License: BSD 3 clause
 
import numpy as np
import matplotlib.pyplot as plt
from sklearn import svm
 
# we create 40 separable points
np.random.seed(0)
X = np.r_[np.random.randn(20, 2) - [2, 2], np.random.randn(20, 2) + [2, 2]]
Y = [0] * 20 + [1] * 20
 
# figure number
fignum = 1
 
# fit the model
for name, penalty in (('unreg', 1), ('reg', 0.05)):
 
    clf = svm.SVC(kernel='linear', C=penalty)
    clf.fit(X, Y)
 
    # get the separating hyperplane
    w = clf.coef_[0]
    a = -w[0] / w[1]
    xx = np.linspace(-5, 5)
    yy = a * xx - (clf.intercept_[0]) / w[1]
 
    # plot the parallels to the separating hyperplane that pass through the
    # support vectors
    margin = 1 / np.sqrt(np.sum(clf.coef_ ** 2))
    yy_down = yy + a * margin
    yy_up = yy - a * margin
 
    # plot the line, the points, and the nearest vectors to the plane
    plt.figure(fignum, figsize=(4, 3))
    plt.clf()
    plt.plot(xx, yy, 'k-')
    plt.plot(xx, yy_down, 'k--')
    plt.plot(xx, yy_up, 'k--')
 
    plt.scatter(clf.support_vectors_[:, 0], clf.support_vectors_[:, 1], s=80,
                facecolors='none', zorder=10)
    plt.scatter(X[:, 0], X[:, 1], c=Y, zorder=10, cmap=plt.cm.Paired)
 
    plt.axis('tight')
    x_min = -4.8
    x_max = 4.2
    y_min = -6
    y_max = 6
 
    XX, YY = np.mgrid[x_min:x_max:200j, y_min:y_max:200j]
    Z = clf.predict(np.c_[XX.ravel(), YY.ravel()])
 
    # Put the result into a color plot
    Z = Z.reshape(XX.shape)
    plt.figure(fignum, figsize=(4, 3))
    plt.pcolormesh(XX, YY, Z, cmap=plt.cm.Paired)
 
    plt.xlim(x_min, x_max)
    plt.ylim(y_min, y_max)
 
    plt.xticks(())
    plt.yticks(())
    fignum = fignum + 1
 
plt.show()

Total running time of the script: (0 minutes 0.147 seconds)

Download Python source code: plot_svm_margin.py

Download IPython notebook: plot_svm_margin.ipynb

Links:

http://scikit-learn.org/stable/auto_examples/svm/plot_svm_margin.html

doc_scikit_learn

2025-01-10 15:47:30

Comments