diff --git a/tree.py b/tree.py
index 0e624aa..1a38696 100644
--- a/tree.py
+++ b/tree.py
@@ -4,6 +4,7 @@
 import operator
 import treePlotter
 from collections import Counter
+from copy import deepcopy
 
 
 pre_pruning = True
@@ -55,10 +56,10 @@ def cal_entropy(dataset):
     labelCounts = {}
     # 给所有可能分类创建字典
     for featVec in dataset:
-        currentlabel = featVec[-1]
+        currentlabel = featVec[-1] 
         if currentlabel not in labelCounts.keys():
             labelCounts[currentlabel] = 0
-        labelCounts[currentlabel] += 1
+        labelCounts[currentlabel] += 1 
     Ent = 0.0
     for key in labelCounts:
         p = float(labelCounts[key]) / numEntries
@@ -184,6 +185,7 @@ def ID3_createTree(dataset, labels, test_dataset):
     print(u"此时最优索引为：" + (bestFeatLabel))
 
     ID3Tree = {bestFeatLabel: {}}
+    labels_for_post_pruning = deepcopy(labels)
     del (labels[bestFeat])
     # 得到列表包括节点所有的属性值
     featValues = [example[bestFeat] for example in dataset]
@@ -224,8 +226,8 @@ def ID3_createTree(dataset, labels, test_dataset):
 
     if post_pruning:
         tree_output = classifytest(ID3Tree,
-                                   featLabels=['年龄段', '有工作', '有自己的房子', '信贷情况'],
-                                   testDataSet=test_dataset)
+                                   featLabels=labels_for_post_pruning,
+                                   testDataSet=test_dataset)# 这里传入的数据集的特征集合是变化后的，所以应该传入变化的特征集
         ans = []
         for vec in test_dataset:
             ans.append(vec[-1])