如何使用sklearn模块中的svm

例子1.

Image [15].png

我们采用上面这个例子：

from sklearn import svm

#x是（2，0），（1，1），（2，3）三个点
x = [[2, 0], [1, 1], [2, 3]]
y = [0, 0, 1]
clf = svm.SVC(kernel = 'linear')

clf.fit(x, y)


print("==========clf===========")
print (clf)
print("=======clf.support_vectors_==============")
# get support vectors
print (clf.support_vectors_)
# get indices of support vectors
print("========clf.support_=============")
print (clf.support_)
# get number of support vectors for each class
print("=======clf.n_support_==============")
print (clf.n_support_)

print("==========predict===========")
#预测点（2，0）

print(clf.predict([[2,0]]))

运行结果：

屏幕快照 2018-09-03 下午8.30.04.png

例子2

使用numpy生成数据
使用pylab画出图形
下面是相应的代码

import numpy as np
import pylab as pl
from sklearn import svm

# we create 40 separable points
np.random.seed(0)#保证每次运行时抓的值不变
X = np.r_[np.random.randn(20, 2) - [2, 2], np.random.randn(20, 2) + [2, 2]]#取20个点维度是2
Y = [0]*20 +[1]*20#前面20个点归为0，后面20个点归为1

#fit the model
clf = svm.SVC(kernel='linear')
clf.fit(X, Y)

# get the separating hyperplane
w = clf.coef_[0]#取的w的值
a = -w[0]/w[1]#点斜式的斜率
xx = np.linspace(-5, 5)#从-5到5产生连续的值
yy = a*xx - (clf.intercept_[0])/w[1]#clf.intercept_[0]相当于是w3

# plot the parallels to the separating hyperplane that pass through the support vectors
b = clf.support_vectors_[0]
yy_down = a*xx + (b[1] - a*b[0])
b = clf.support_vectors_[-1]#-1指的是最后一个值
yy_up = a*xx + (b[1] - a*b[0])

print( "w: ")
print(w)
print ("a: ")
print(a)
# print "xx: ", xx
# print "yy: ", yy
print ("support_vectors_: ")
print("clf.support_vectors_")
print ("clf.coef_: ")
print(clf.coef_)

# switching to the generic n-dimensional parameterization of the hyperplan to the 2D-specific equation
# of a line y=a.x +b: the generic w_0x + w_1y +w_3=0 can be rewritten y = -(w_0/w_1) x + (w_3/w_1)


# plot the line, the points, and the nearest vectors to the plane
pl.plot(xx, yy, 'k-')
pl.plot(xx, yy_down, 'k--')
pl.plot(xx, yy_up, 'k--')

pl.scatter(clf.support_vectors_[:, 0], clf.support_vectors_[:, 1],
          s=80, facecolors='none')
pl.scatter(X[:, 0], X[:, 1], c=Y, cmap=pl.cm.Paired)

pl.axis('tight')
pl.show()

运行结果：

屏幕快照 2018-09-03 下午8.33.55.png
Figure_1.png