決策樹(Decision Tree）是在已知各種情況發生概率的基礎上，通過構成決策樹來求取淨現值的期望值大於等於零的概率，評價專案風險，判斷其可行性的決策分析方法，是直觀運用概率分析的一種圖解法。一些基礎原理這裡就不再一一介紹了，直接進入今天的主題，如何視覺化決策樹。

本篇使用klearn來實現決策樹的過程，下面是詳細講解：

首先匯入必要的包：

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from sklearn.model_selection import GridSearchCV
from sklearn.metrics import accuracy_score
from sklearn.metrics import roc_auc_score
 

data = pd.read_csv("mushrooms.csv")
data.head()

可以看出這是一個分類變數的資料集。然後，我們就要將它變成數值變數，好利於下面的建模。

from sklearn.preprocessing import LabelEncoder
labelencoder = LabelEncoder()
for col in data.columns:
    data[col] = labelencoder.fit_transform(data[col])
data.head()
 

data.shape

y = data['class']
X = data.drop('class', axis=1)
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0, train_size=0.8)
columns = X_train.columns

# 資料標準化
from sklearn.preprocessing import StandardScaler
ss_X = StandardScaler()
ss_y = StandardScaler()
X_train = ss_X.fit_transform(X_train)
X_test = ss_X.transform(X_test)
 

from sklearn.tree import DecisionTreeClassifier
model_tree = DecisionTreeClassifier()
model_tree.fit(X_train, y_train)

y_prob = model_tree.predict_proba(X_test)[:,1]
y_pred = np.where(y_prob > 0.5, 1, 0)
model_tree.score(X_test, y_pred)

說明決策樹非常吻合此資料集。

最後，完成決策樹的視覺化

# 視覺化樹圖
data_ = pd.read_csv("mushrooms.csv")
data_feature_name = data_.columns[1:]
data_target_name = np.unique(data_["class"])
import graphviz
import pydotplus
from sklearn import tree
from IPython.display import Image
import os
os.environ["PATH"] += os.pathsep + 'C:/Program Files (x86)/Graphviz2.38/bin/'
dot_tree = tree.export_graphviz(model_tree,out_file=None,feature_names=data_feature_name,class_names=data_target_name,filled=True, rounded=True,special_characters=True)
graph = pydotplus.graph_from_dot_data(dot_tree)
img = Image(graph.create_png())
graph.write_png("out.png")

注意：graphviz包不僅需要使用pip install graphviz安裝還需要單獨安裝。使用時，還需要引入graphviz絕對路徑。

參考：http://scikit-learn.org/stable/modules/tree.html

graphviz-2.38.msi安裝包下載：http://www.graphviz.org/Download_windows.php

資料集：http://download.csdn.net/download/llh_1178/10115766