机器学习多分类和多标签处理方法

#coding=utf-8
from sklearn import metrics
from sklearn import cross_validation
from sklearn.svm import SVC
from sklearn.multiclass import OneVsRestClassifier
from sklearn.preprocessing import MultiLabelBinarizer
import numpy as np
from numpy import random
X=np.arange(15).reshape(5,3)
y=np.arange(5)
Y_1 = np.arange(5)
random.shuffle(Y_1)
Y_2 = np.arange(5)
random.shuffle(Y_2)
Y = np.c_[Y_1,Y_2]
def multiclassSVM():
  X_train, X_test, y_train, y_test = cross_validation.train_test_split(X, y, test_size=0.2,random_state=0)
  model = OneVsRestClassifier(SVC())
  model.fit(X_train, y_train)
  predicted = model.predict(X_test)
  print predicted
def multilabelSVM():
  Y_enc = MultiLabelBinarizer().fit_transform(Y)
  X_train, X_test, Y_train, Y_test = cross_validation.train_test_split(X, Y_enc, test_size=0.2, random_state=0)
  model = OneVsRestClassifier(SVC())
  model.fit(X_train, Y_train)
  predicted = model.predict(X_test)
  print predicted
if __name__ == '__main__':
  multiclassSVM()
  # multilabelSVM()

作者：Cer_ml
链接：http://www.jianshu.com/p/516f009c0875
來源：简书
著作权归作者所有。商业转载请联系作者获得授权，非商业转载请注明出处。