《深度学习Python实践》附录——决策树模型可视化

1：图形可视化需要安装：Graphviz

ubuntu下安装Graphviz：https://blog.csdn.net/caiandyong/article/details/44408831
1）sudo apt-get install graphviz, graphviz-dev
2） sudo pip3 install pydot

2：scikit-learn 的算法选择路径

一个网站：http://sklearn.lzjqsdd.com/

原文链接：https://blog.csdn.net/yimingsilence/article/details/70160602?locationNum=7&fps=1

---

3.决策树模型可视化：

from pandas import read_csv
from matplotlib import pyplot
from matplotlib.image import imread
from sklearn.model_selection import train_test_split
from sklearn.tree import DecisionTreeClassifier
from sklearn.tree import export_graphviz
from sklearn.metrics import accuracy_score
import pydotplus
import os

filename='iris.txt'
names=['separ-length','separ-width','petal-length','petal-width','class']
dataset=read_csv(filename,names=names)

#分离数据集
#80%用于训练，20%用于评估
array=dataset.values
X=array[:,0:4]
Y=array[:,4]
validation_size=0.2
seed=7
X_train,X_validation,Y_train,Y_validation=\
    train_test_split(X,Y,test_size=validation_size,random_state=seed)


#算法审查

model=DecisionTreeClassifier()
model.fit(X=X_train,y=Y_train)

#决策树图形化
dot_data=export_graphviz(model,out_file=None)
graph=pydotplus.graph_from_dot_data(dot_data)

path=os.getcwd() + '/'
tree_file=path+'iris.png'


try:
    os.remove(tree_file)
except:
    print('there is no file to be deleted.')
finally:
    graph.write(tree_file,format = 'png')

#显示图像
image_data=imread(tree_file)
pyplot.imshow(image_data)
pyplot.axis('off')
pyplot.show()

#评估算法
predictions=model.predict(X_validation)
print(accuracy_score(Y_validation,predictions))

运行结果：

0.8666666666666667