因为最近实习的需要，所以用python里的sklearn包重新写了一次决策树

工具：sklearn，http://www.lfd.uci.edu/~gohlke/pythonlibs/#numpy；将dot文件转化为pdf格式（是为了将形成的决策树可视化）graphviz-2.38，下载解压之后将其中的bin文件的目录添加进环境变量

源代码如下：

from sklearn.feature_extraction import DictVectorizer
import csv
from sklearn import tree
from sklearn import preprocessing
from sklearn.externals.six import StringIO
from xml.sax.handler import feature_external_ges
from numpy.distutils.fcompiler import dummy_fortran_file

# Read in the csv file and put features into list of dict and list of class label
allElectronicsData = open(r'E:/DeepLearning/resources/AllElectronics.csv', 'rt')
reader = csv.reader(allElectronicsData)
headers = next(reader)
featureList = []
lableList = []
for row in reader:
lableList.append(row[len(row)-1])
rowDict = {}
#不包括len(row)-1
for i in range(1,len(row)-1):
rowDict[headers[i]] = row[i]
featureList.append(rowDict)
print(featureList)

vec = DictVectorizer()
dummX = vec.fit_transform(featureList).toarray()
print(str(dummX))
lb = preprocessing.LabelBinarizer()
dummY = lb.fit_transform(lableList)
print(str(dummY))

#entropy=>ID3
clf = tree.DecisionTreeClassifier(criterion='entropy')
clf = clf.fit(dummX, dummY)
print("clf:"+str(clf))

#可视化tree
with open("resultTree.dot",'w')as f:
f = tree.export_graphviz(clf, feature_names=vec.get_feature_names(),out_file = f)

#对于新的数据怎样来查看它的分类
oneRowX = dummX[0,:]
print("oneRowX: "+str(oneRowX))
newRowX = oneRowX
newRowX[0] = 1
newRowX[2] = 0

predictedY = clf.predict(newRowX)
print("predictedY: "+ str(predictedY))

这里的AllElectronics.csv，形式如下图所示：

今天早上好不容易将jdk、eclipse以及pydev装进linux，但是，但是，但是，想装numpy的时候，总是报错，发现是没有gcc，然后又去装gcc，真是醉了，到现在gcc还是没有装成功，再想想法子，实在不行过段时间去问大神