# -*- coding: utf-8 -*-
"""
Created on Sat Nov 9 10:42:38 2019

@author: asus
"""
"""
決策樹
目的：
1. 使用決策樹模型
2. 了解決策樹模型的參數(shù)
3. 初步了解調(diào)參數(shù)
要求：
基于乳腺癌數(shù)據(jù)集完成以下任務(wù)：
1.調(diào)整參數(shù)criterion，使用不同算法信息熵（entropy）和基尼不純度算法(gini)
2.調(diào)整max_depth參數(shù)值，查看不同的精度
3.根據(jù)參數(shù)criterion和max_depth得出你初步的結(jié)論。
"""

import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import mglearn 
from sklearn.model_selection import train_test_split
#導(dǎo)入乳腺癌數(shù)據(jù)集
from sklearn.datasets import load_breast_cancer
from sklearn.tree import DecisionTreeClassifier


#決策樹并非深度越大越好，考慮過擬合的問題
#mglearn.plots.plot_animal_tree()
#mglearn.plots.plot_tree_progressive()

#獲取數(shù)據(jù)集
cancer = load_breast_cancer()
#對數(shù)據(jù)集進(jìn)行切片
X_train,X_test,y_train,y_test = train_test_split(cancer.data,cancer.target,
       stratify = cancer.target,random_state = 42)
#查看訓(xùn)練集和測試集數(shù)據(jù)      
print('train dataset :{0} ;test dataset :{1}'.format(X_train.shape,X_test.shape))
#建立模型(基尼不純度算法(gini)),使用不同最大深度和隨機(jī)狀態(tài)和不同的算法看模型評(píng)分
tree = DecisionTreeClassifier(random_state = 0,criterion = 'gini',max_depth = 5)
#訓(xùn)練模型
tree.fit(X_train,y_train)
#評(píng)估模型
print("Accuracy(準(zhǔn)確性) on training set: {:.3f}".format(tree.score(X_train, y_train)))
print("Accuracy(準(zhǔn)確性) on test set: {:.3f}".format(tree.score(X_test, y_test)))
print(tree)


# 參數(shù)選擇 max_depth,算法選擇基尼不純度算法(gini) or 信息熵（entropy）
def Tree_score(depth = 3,criterion = 'entropy'):
 """
 參數(shù)為max_depth(默認(rèn)為3)和criterion(默認(rèn)為信息熵entropy)，
 函數(shù)返回模型的訓(xùn)練精度和測試精度
 """
 tree = DecisionTreeClassifier(criterion = criterion,max_depth = depth)
 tree.fit(X_train,y_train)
 train_score = tree.score(X_train, y_train)
 test_score = tree.score(X_test, y_test)
 return (train_score,test_score)

#gini算法，深度對模型精度的影響
depths = range(2,25)#考慮到數(shù)據(jù)集有30個(gè)屬性
scores = [Tree_score(d,'gini') for d in depths]
train_scores = [s[0] for s in scores]
test_scores = [s[1] for s in scores]

plt.figure(figsize = (6,6),dpi = 144)
plt.grid()
plt.xlabel("max_depth of decision Tree")
plt.ylabel("score")
plt.title("'gini'")
plt.plot(depths,train_scores,'.g-',label = 'training score')
plt.plot(depths,test_scores,'.r--',label = 'testing score')
plt.legend()


#信息熵（entropy），深度對模型精度的影響
scores = [Tree_score(d) for d in depths]
train_scores = [s[0] for s in scores]
test_scores = [s[1] for s in scores]

plt.figure(figsize = (6,6),dpi = 144)
plt.grid()
plt.xlabel("max_depth of decision Tree")
plt.ylabel("score")
plt.title("'entropy'")
plt.plot(depths,train_scores,'.g-',label = 'training score')
plt.plot(depths,test_scores,'.r--',label = 'testing score')
plt.legend()