diff --git a/src/python/03.DecisionTree/DTSklearn.py b/src/python/03.DecisionTree/DTSklearn.py
new file mode 100644
index 00000000..5155b214
--- /dev/null
+++ b/src/python/03.DecisionTree/DTSklearn.py
@@ -0,0 +1,115 @@
+#!/usr/bin/python
+# coding: utf8
+# 原始链接： http://blog.csdn.net/lsldd/article/details/41223147
+import numpy as np
+from sklearn import tree
+from sklearn.metrics import precision_recall_curve
+from sklearn.metrics import classification_report
+from sklearn.cross_validation import train_test_split
+
+
+def createDataSet():
+    ''' 数据读入 '''
+    data = []
+    labels = []
+    with open("testData/DT_data.txt") as ifile:
+        for line in ifile:
+            # 特征： 身高 体重   label： 胖瘦
+            tokens = line.strip().split(' ')
+            data.append([float(tk) for tk in tokens[:-1]])
+            labels.append(tokens[-1])
+    # 特征数据
+    x = np.array(data)
+    # label分类的标签数据
+    labels = np.array(labels)
+    # 预估结果的标签数据
+    y = np.zeros(labels.shape)
+
+    ''' 标签转换为0/1 '''
+    y[labels == 'fat'] = 1
+    print data, '-------', x, '-------', labels, '-------', y
+    return x, y
+
+
+def predict_train(x_train, y_train):
+    '''
+    使用信息熵作为划分标准，对决策树进行训练
+    参考链接： http://scikit-learn.org/stable/modules/generated/sklearn.tree.DecisionTreeClassifier.html#sklearn.tree.DecisionTreeClassifier
+    '''
+    clf = tree.DecisionTreeClassifier(criterion='entropy')
+    # print(clf)
+    clf.fit(x_train, y_train)
+    ''' 系数反映每个特征的影响力。越大表示该特征在分类中起到的作用越大 '''
+    print 'feature_importances_: %s' % clf.feature_importances_
+
+    '''测试结果的打印'''
+    y_pre = clf.predict(x_train)
+    # print(x_train)
+    print(y_pre)
+    print(y_train)
+    print(np.mean(y_pre == y_train))
+    return y_pre, clf
+
+
+def show_precision_recall(x, clf,  y_train, y_pre):
+    '''
+    准确率与召回率
+    参考链接： http://scikit-learn.org/stable/modules/generated/sklearn.metrics.precision_recall_curve.html#sklearn.metrics.precision_recall_curve
+    '''
+    precision, recall, thresholds = precision_recall_curve(y_train, y_pre)
+    # 计算全量的预估结果
+    answer = clf.predict_proba(x)[:, 1]
+
+    '''
+    展现 准确率与召回率
+        precision 准确率
+        recall 召回率
+        f1-score  准确率和召回率的一个综合得分
+        support 参与比较的数量
+    参考链接：http://scikit-learn.org/stable/modules/generated/sklearn.metrics.classification_report.html#sklearn.metrics.classification_report
+    '''
+    # target_names 以 y的label分类为准
+    target_names = ['thin', 'fat']
+    print(classification_report(y, answer, target_names=target_names))
+    print(answer)
+    print(y)
+
+
+def show_pdf(clf):
+    '''
+    可视化输出
+    把决策树结构写入文件: http://sklearn.lzjqsdd.com/modules/tree.html
+
+    Mac报错：pydotplus.graphviz.InvocationException: GraphViz's executables not found
+    解决方案：sudo brew install graphviz
+    参考写入： http://www.jianshu.com/p/59b510bafb4d
+    '''
+    # with open("testResult/tree.dot", 'w') as f:
+    #     from sklearn.externals.six import StringIO
+    #     tree.export_graphviz(clf, out_file=f)
+
+    import pydotplus
+    from sklearn.externals.six import StringIO
+    dot_data = StringIO()
+    tree.export_graphviz(clf, out_file=dot_data)
+    graph = pydotplus.graph_from_dot_data(dot_data.getvalue())
+    graph.write_pdf("testResult/tree.pdf")
+
+    # from IPython.display import Image
+    # Image(graph.create_png())
+
+if __name__ == '__main__':
+    x, y = createDataSet()
+
+    ''' 拆分训练数据与测试数据， 80%做训练 20%做测试 '''
+    x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.2)
+    print '拆分数据：', x_train, x_test, y_train, y_test
+
+    # 得到训练的预测结果集
+    y_pre, clf = predict_train(x_train, y_train)
+
+    # 展现 准确率与召回率
+    show_precision_recall(x, clf, y_train, y_pre)
+
+    # 可视化输出
+    show_pdf(clf)
diff --git a/src/python/03.DecisionTree/DecisionTree.py b/src/python/03.DecisionTree/DecisionTree.py
index c7920bd1..03a4d25f 100644
--- a/src/python/03.DecisionTree/DecisionTree.py
+++ b/src/python/03.DecisionTree/DecisionTree.py
@@ -129,7 +129,7 @@ def chooseBestFeatureToSplit(dataSet):
             subDataSet = splitDataSet(dataSet, i, value)
             prob = len(subDataSet)/float(len(dataSet))
             newEntropy += prob * calcShannonEnt(subDataSet)
-        # 计算label的信息熵和每个特征的信息熵 的增益值，如果增益值大于最大值，那么效果越好
+        # gain[信息增益] 值越大，意味着该分类提供的信息量越大，该特征对分类的不确定程度越小
         infoGain = baseEntropy - newEntropy
         if (infoGain > bestInfoGain):
             bestInfoGain = infoGain
diff --git a/src/python/tools/DecisionTree.py b/src/python/tools/DecisionTree.py
new file mode 100644
index 00000000..a2974d91
--- /dev/null
+++ b/src/python/tools/DecisionTree.py
@@ -0,0 +1,124 @@
+#!/usr/bin/python
+# coding: utf8
+
+from math import log
+
+
+def calcShannonEnt(dataSet):
+    """calcShannonEnt(calculate Shannon entropy 计算label分类标签的香农熵)
+
+    Args:
+        dataSet 数据集
+    Returns:
+        返回香农熵的计算值
+    Raises:
+
+    """
+    # 求list的长度，表示计算参与训练的数据量
+    numEntries = len(dataSet)
+    # print type(dataSet), 'numEntries: ', numEntries
+
+    # 计算分类标签label出现的次数
+    labelCounts = {}
+    # the the number of unique elements and their occurance
+    for featVec in dataSet:
+        currentLabel = featVec[-1]
+        if currentLabel not in labelCounts.keys():
+            labelCounts[currentLabel] = 0
+        labelCounts[currentLabel] += 1
+        # print '-----', featVec, labelCounts
+
+    # 对于label标签的占比，求出label标签的香农熵
+    shannonEnt = 0.0
+    for key in labelCounts:
+        prob = float(labelCounts[key])/numEntries
+        # log base 2
+        shannonEnt -= prob * log(prob, 2)
+        # print '---', prob, prob * log(prob, 2), shannonEnt
+    return shannonEnt
+
+
+def splitDataSet(dataSet, axis, value):
+    """splitDataSet(通过遍历dataSet数据集，求出axis对应的colnum列的值为value的行)
+
+    Args:
+        dataSet 数据集
+        axis 表示每一行的axis列
+        value 表示axis列对应的value值
+    Returns:
+        axis列为value的数据集【该数据集需要排除axis列】
+    Raises:
+
+    """
+    retDataSet = []
+    for featVec in dataSet:
+        # axis列为value的数据集【该数据集需要排除axis列】
+        if featVec[axis] == value:
+            # chop out axis used for splitting
+            reducedFeatVec = featVec[:axis]
+            '''
+            请百度查询一下： extend和append的区别
+            '''
+            reducedFeatVec.extend(featVec[axis+1:])
+            # 收集结果值 axis列为value的行【该行需要排除axis列】
+            retDataSet.append(reducedFeatVec)
+    return retDataSet
+
+
+def getFeatureShannonEnt(dataSet, labels):
+    """chooseBestFeatureToSplit(选择最好的特征)
+
+    Args:
+        dataSet 数据集
+    Returns:
+        bestFeature 最优的特征列
+    Raises:
+
+    """
+    # 求第一行有多少列的 Feature
+    numFeatures = len(dataSet[0]) - 1
+    # label的信息熵
+    baseEntropy = calcShannonEnt(dataSet)
+    # 最优的信息增益值, 和最优的Featurn编号
+    bestInfoGain, bestFeature, endEntropy = 0.0, -1, 0.0
+    # iterate over all the features
+    for i in range(numFeatures):
+        # create a list of all the examples of this feature
+        # 获取每一个feature的list集合
+        featList = [example[i] for example in dataSet]
+        # get a set of unique values
+        # 获取剔重后的集合
+        uniqueVals = set(featList)
+        # 创建一个临时的信息熵
+        newEntropy = 0.0
+        # 遍历某一列的value集合，计算该列的信息熵
+        for value in uniqueVals:
+            subDataSet = splitDataSet(dataSet, i, value)
+            prob = len(subDataSet)/float(len(dataSet))
+            newEntropy += prob * calcShannonEnt(subDataSet)
+        # gain[信息增益] 值越大，意味着该分类提供的信息量越大，该特征对分类的不确定程度越小
+        # gain[信息增益]=0, 表示与类别相同，无需其他的分类
+        # gain[信息增益]=baseEntropy, 表示分类和没分类没有区别
+        infoGain = baseEntropy - newEntropy
+        # print infoGain
+        if (infoGain > bestInfoGain):
+            endEntropy = newEntropy
+            bestInfoGain = infoGain
+            bestFeature = i
+    else:
+        if numFeatures < 0:
+            labels[bestFeature] = 'null'
+
+    return labels[bestFeature], baseEntropy, endEntropy, bestInfoGain
+
+
+if __name__ == '__main__':
+    labels = ['no surfacing', 'flippers']
+    dataSet1 = [['yes'], ['yes'], ['no'], ['no'], ['no']]
+    dataSet2 = [['a', 1, 'yes'], ['a', 2, 'yes'], ['b', 3, 'no'], ['c', 4, 'no'], ['c', 5, 'no']]
+    dataSet3 = [[1, 'yes'], [1, 'yes'], [1, 'no'], [3, 'no'], [3, 'no']]
+    infoGain1 = getFeatureShannonEnt(dataSet1, labels)
+    infoGain2 = getFeatureShannonEnt(dataSet2, labels)
+    infoGain3 = getFeatureShannonEnt(dataSet3, labels)
+    print '香农熵: \n\t%s, \n\t%s, \n\t%s' % (infoGain1, infoGain2, infoGain3)
+
diff --git a/testData/DT_data.txt b/testData/DT_data.txt
new file mode 100644
index 00000000..3bf55b88
--- /dev/null
+++ b/testData/DT_data.txt
@@ -0,0 +1,10 @@
+1.5 50 thin  
+1.5 60 fat  
+1.6 40 thin  
+1.6 60 fat  
+1.7 60 thin  
+1.7 80 fat  
+1.8 60 thin  
+1.8 90 fat  
+1.9 70 thin  
+1.9 80 thin 
diff --git a/testResult/tree.pdf b/testResult/tree.pdf
new file mode 100644
index 00000000..f2b72829
Binary files /dev/null and b/testResult/tree.pdf differ