1 import numpy as np 2 import matplotlib.pyplot as plt 3 from sklearn import svm 4 from sklearn.datasets import make_blobs 5 X,y=make_blobs(n_samples=50,centers=2,random_state=6) 6 clf=svm.SVC(kernel='linear',C=1000) 7 #clf=svm.SVC(kernel='rbf',C=1000) 8 clf.fit(X,y) 9 print("the score of this model:{}".format(clf.score(X,y))) 10 print(clf.predict(X)) 11 12 plt.scatter(X[:,0],X[:,1],c=y,s=30,cmap=plt.cm.Paired) 13 ax=plt.gca() 14 xlim=ax.get_xlim() 15 ylim=ax.get_ylim() 16 17 xx=np.linspace(xlim[0],xlim[1],30) 18 yy=np.linspace(ylim[0],ylim[1],30) 19 YY,XX=np.meshgrid(yy,xx) 20 xy=np.vstack([XX.ravel(),YY.ravel()]).T 21 z=clf.decision_function(xy).reshape(XX.shape) 22 23 ax.contour(XX,YY,z,colors='k',levels=[-1,0,1],alpha=0.5,linestyles=['--','-','--']) 24 ax.scatter(clf.support_vectors_[:,0],clf.support_vectors_[:,1],s=100,linewidths=1,facecolors='none') 25 plt.show()
1 X,y=make_blobs(n_samples=50,centers=2,random_state=6) 2 #clf=svm.SVC(kernel='linear',C=1000) 3 clf=svm.SVC(kernel='rbf',C=1000) 4 clf.fit(X,y) 5 print("the score of this model:{}".format(clf.score(X,y))) 6 print(clf.predict(X)) 7 8 plt.scatter(X[:,0],X[:,1],c=y,s=30,cmap=plt.cm.Paired) 9 ax=plt.gca() 10 xlim=ax.get_xlim() 11 ylim=ax.get_ylim() 12 13 xx=np.linspace(xlim[0],xlim[1],30) 14 yy=np.linspace(ylim[0],ylim[1],30) 15 YY,XX=np.meshgrid(yy,xx) 16 xy=np.vstack([XX.ravel(),YY.ravel()]).T 17 z=clf.decision_function(xy).reshape(XX.shape) 18 19 ax.contour(XX,YY,z,colors='k',levels=[-1,0,1],alpha=0.5,linestyles=['--','-','--']) 20 ax.scatter(clf.support_vectors_[:,0],clf.support_vectors_[:,1],s=100,linewidths=1,facecolors='none') 21 plt.show()
1 import numpy as np 2 def make_meshgrid(x,y,h=.02): 3 x_min = x.min() - 1 4 x_max = x.max() + 1 5 y_min = y.min() - 1 6 y_max = y.max() + 1 7 xx,yy=np.meshgrid(np.arange(x_min,x_max,h), 8 np.arange(y_min,y_max,h)) 9 return xx,yy 10 11 def plot_contours(ax,clf,xx,yy,**params): 12 z=clf.predict(np.c_[xx.ravel(),yy.ravel()]) 13 z=z.reshape(xx.shape) 14 out=ax.contourf(xx,yy,z,**params) 15 return out 16 17 from sklearn.datasets import load_wine 18 wine=load_wine() 19 X=wine.data[:,:2] 20 y=wine.target 21 C=1.0 22 23 from sklearn import svm 24 import matplotlib.pyplot as plt 25 models=(svm.SVC(kernel='linear',C=C), 26 svm.LinearSVC(C=C,max_iter=10000), 27 svm.SVC(kernel='rbf',gamma=0.7,C=C), 28 svm.SVC(kernel='poly',degree=3,C=C) 29 ) 30 models=(clf.fit(X,y) for clf in models) 31 titles=('SVC with linear kernal L1', 32 'LinearSVC(linear kernal) L2', 33 'SVC with RBF kernal', 34 'SVC with polynomial(degree=3) kernal' 35 ) 36 fig,sub=plt.subplots(2,2) 37 plt.subplots_adjust(wspace=0.4,hspace=0.4) 38 X0,X1=X[:,0],X[:,1] 39 xx,yy=make_meshgrid(X0,X1) 40 for clf,title,ax in zip(models,titles,sub.flatten()): 41 plot_contours(ax,clf,xx,yy,cmap=plt.cm.plasma,alpha=0.8) 42 ax.scatter(X0,X1,c=y,cmap=plt.cm.plasma,s=20,edgecolors='k') 43 ax.set_xlim(xx.min(),xx.max()) 44 ax.set_ylim(yy.min(),yy.max()) 45 ax.set_xlabel("Feature 0") 46 ax.set_ylabel("Feature 1") 47 ax.set_xticks(()) 48 ax.set_yticks(()) 49 ax.set_title(title) 50 plt.show()
1 models=(svm.SVC(kernel='rbf',gamma=0.1,C=C), 2 svm.SVC(kernel='rbf',gamma=1,C=C), 3 svm.SVC(kernel='rbf',gamma=10,C=C), 4 ) 5 models=(clf.fit(X,y) for clf in models) 6 titles=('RBF-kernal SVC gamma=0.1', 7 'RBF-kernal SVC gamma=1', 8 'RBF-kernal SVC gamma=10' 9 ) 10 fig,sub=plt.subplots(1,3,figsize=(15,3)) 11 12 X0,X1=X[:,0],X[:,1] 13 xx,yy=make_meshgrid(X0,X1) 14 for clf,title,ax in zip(models,titles,sub.flatten()): 15 plot_contours(ax,clf,xx,yy,cmap=plt.cm.plasma,alpha=0.8) 16 ax.scatter(X0,X1,c=y,cmap=plt.cm.plasma,s=20,edgecolors='k') 17 ax.set_xlim(xx.min(),xx.max()) 18 ax.set_ylim(yy.min(),yy.max()) 19 ax.set_xlabel("Feature 0") 20 ax.set_ylabel("Feature 1") 21 ax.set_xticks(()) 22 ax.set_yticks(()) 23 ax.set_title(title) 24 plt.show()