# -*- coding: utf-8 -*-
"""
Created on Fri Jul 19 14:47:14 2019
@author: Caiyunbin
"""
import pandas as pd
import numpy as np
from sklearn import tree
from sklearn.tree import DecisionTreeClassifier
import graphviz
#定义一个数据集或者说导入一个数据集
def create_data():
row_data = {'no surfacing':[1,1,1,0,0],
'flippers':[1,1,0,1,1],
'fish':['yes','yes','no','no','no']}
dataset = pd.DataFrame(row_data)
return dataset
#定义香浓熵的算法
def calent(dataset):
n = dataset.shape[0] #注意shape的中括号
nclass = dataset.iloc[:,-1].value_counts()
p = nclass/n
ent = (-p*np.log2(p)).sum()
return ent
#定义最佳分割的列方法
def best_split(dataset):
baseent = calent(dataset)
bestgain = 0
axis = -1
for i in range(dataset.shape[1]-1):
levels = dataset.iloc[:,i].value_counts().index
ents = 0
for j in levels:
childset = dataset[dataset.iloc[:,i]==j]
ent = calent(childset)
ents += (childset.shape[0]/dataset.shape[0])*ent
infogain = baseent - ents
if (infogain>bestgain):
bestgain = infogain
axis = i
return axis
#定义一个划分数据集的方法
def mysplit(dataset,axis,value):
col = dataset.columns[axis]
redataset = dataset.loc[dataset[col]==value,:].drop(col,axis=1)
return redataset
#使用递归的方法构建一棵树
def createtree(dataset):
colum_all_feature = list(dataset.columns)
classlist = dataset.iloc[:,-1].value_counts() #会按照从高到低的顺序进行排列
if classlist[0] == dataset.shape[0] or dataset.shape[1]==1:
return classlist.index[0]
axis = best_split(dataset)
bestfeature = colum_all_feature[axis]
mytree = {bestfeature:{}}
del colum_all_feature[axis]
valuelist = set(dataset.iloc[:,axis])
for value in valuelist:
mytree[bestfeature][value] = createtree(mysplit(dataset,axis,value))
return mytree
#关于决策树的存储，格式为numpy库npy文件
def save_tree(mytree):
np.save('mytree.npy',mytree)
#将数在测试集上进行测试
def classy(inputtree,labels,testvec): #判断其中的一条向量
firststr = next(iter(inputtree))
seconddict = inputtree[firststr]
featindex = labels.index(firststr)
for key in seconddict.keys():
if testvec[featindex]==key:
if type(seconddict[key])==dict:
classlabel = classy(seconddict[key],labels,testvec)
else:
classlabel = seconddict[key]
return classlabel
#查看该模型的性能
def acc_classify(train,test):
inputtree = createtree(train)
labels = list(train.columns)
result = []
for i in range(test.shape[0]):
testvec = test.iloc[i,:-1]
classlabel = classy(inputtree,labels,testvec)
result.append(classlabel)
test['predict']=result
acc = (test.iloc[:,-1]==test.iloc[:-2]).mean()
print('模型的准确率为',acc)
#对决策树进行画图
xtrain = dataset.iloc[:,:-1]
Ytrain = dataset.iloc[:,-1]
labels = Ytrain.unique().tolist
Ytrain = Ytrain.apply(lambda x:labels.index(x))
clf = DecisionTreeClassifier()
clf = clf.fit(xtrain,Ytrain)
tree.export_graphviz(clf)
dot_data = tree.export_graphviz(clf,out_file = None,
feature_names =['no surfacing','flippers'],
class_names=['fish','not fish'],
filled = True,rounded = True,
special_characters = True,
graphviz.source(dot_data))
def main():
dat = create_data()
mytree = createtree(dat)
print(mytree)
if __name__ == '__main__':
main()
mytree=createtree(dataset)
inputtree = mytree
