注意：变量x包含30个5维特征向量元组 . 这些x的值被转移到x_train.x可以想象成x = [[1.0,2.0,3,0,4.0,5.0]，[11.0,12.0,13.0,14.0,15.0]，[21.0， 22.0,23.0,24.0,25,0]，.. .. ..]和y =标签= [1,1,1,2,2,2,3,3,3 ...]我希望申请PCA在x上并减少到两个维度，然后绘制决策边界 . 我可以绘制积分，但无法绘制决策边界
x_train = x
y_train =labels
pca = PCA(n_components=2).fit(x_train)
pca_2d = pca.transform(x_train)
clf = svm.SVC(kernel='linear',C = 3)
clf.fit(pca_2d, y_train)
for i in range(1, pca_2d.shape[0]):
if y_train[i] == 1:
c1 = pl.scatter(pca_2d[i,0],pca_2d[i,1],c='r', s=50,marker='+')
elif y_train[i] == 2:
c2 = pl.scatter(pca_2d[i,0],pca_2d[i,1],c='r', s=50,marker='.')
elif y_train[i] == 3:
c3 = pl.scatter(pca_2d[i,0],pca_2d[i,1],c='r', s=50,marker=',')
elif y_train[i] == 4:
c4 = pl.scatter(pca_2d[i,0],pca_2d[i,1],c='r', s=50,marker='^')
elif y_train[i] == 5:
c5 = pl.scatter(pca_2d[i,0],pca_2d[i,1],c='r', s=50,marker='v')
elif y_train[i] == 6:
x_min, x_max = pca_2d[:, 0].min() - 1, pca_2d[:,0].max() + 1
y_min, y_max = pca_2d[:, 1].min() - 1, pca_2d[:, 1].max() + 1
xx, yy = np.meshgrid(np.arange(x_min, x_max, .01),np.arange(y_min,y_max, .01))
#************ ERROR ******#
Z = clf.predict(np.c_[xx.ravel(), yy.ravel()])
#************ ERROR ******#
Z = Z.reshape(xx.shape)
plt.contourf(xx, yy, Z, cmap=plt.cm.Paired, alpha=0.8)
pl.title('Support Vector Machine Decision Surface')
pl.axis('off')
pl.show()
## The error shown is :
Traceback (most recent call last):
File "D:\New folder_previous.2 - Copy.right\main_pos.py", line 354, in <module>
Z = clf.predict(np.c_[xx.ravel(), yy.ravel()])
File "C:\Python27\lib\site-packages\numpy\lib\index_tricks.py", line 338, in __getitem__
res = _nx.concatenate(tuple(objs), axis=self.axis)
MemoryError
1 回答
错误非常严重，你试图分配太多内存 . 使用大于0.1的步长，研究矩阵的大小，可能是你用xx，yy生成了一个大的 . 此外 - 为什么循环中的所有内容都超过样本？你好像在说这30次，这似乎不是一个合理的想法 .