决策树（decision tree）

尝试了一波单一决策树模型和集成决策树，也体会了一波大量数据训练和增加迭代次数对模型的影响，100,0000棵树了解一下🤣

!/usr/bin/env python
 coding: utf-8
 In[1]:
 from sklearn.model_selection import train_test_split
 In[2]:
 from sklearn.ensemble import GradientBoostingClassifier
 In[3]:
 from sklearn.datasets import load_breast_cancer
 In[4]:
 import matplotlib.pyplot as plt
 In[5]:
 import numpy as np
 In[6]:
 cancer = load_breast_cancer()
 In[7]:
 X_train,X_test,y_train,y_test = train_test_split(cancer.data,cancer.target,random_state = 0)
 In[8]:
 '''
 grbt = GradientBoostingClassifier(random_state = 0,max_depth = 3,learning_rate = 0.01).fit(X_train,y_train)
 '''
 In[9]:
 grbt = GradientBoostingClassifier(random_state = 0,max_depth = 3,n_estimators = 100,learning_rate = 0.1).fit(X_train,y_train)
 In[10]:
 print ('{}'.format(grbt.score(X_train,y_train)))
 In[11]:
 print ('{}'.format(grbt.score(X_test,y_test)))
 In[12]:
 '''
 plot_feature_importances_cancer(grbt)
 '''
 In[13]:
 '''
 def plot_feature_importances_cancer(model):
     n_features = cancer.data.shape()
     plt.barh(range(n_features,model.feature_importance,align = 'center'))
     plt.yticks(np.arange(n_features),cancer.feature_names)
     plt.xlabel('Features importance')
     plt.ylabel('Feature')
 '''
 In[14]:
 n_features = cancer.data.shape[1]
 plt.barh(range(n_features),grbt.feature_importances_,align = 'center')
 plt.yticks(np.arange(n_features),cancer.feature_names)
 plt.xlabel('Features importance')
 plt.ylabel('Feature')
 In[ ]:

grbt = GradientBoostingClassifier(random_state = 0,max_depth = 1,n_estimators = 1000000,learning_rate = 0.1).fit(X_train,y_train)

跑这波代码的时候，笔记本的风扇声音嗡嗡响，看来干这行很烧资源，浅尝辄止，🐸，处理好我的uC方向就好了，深入的没啥兴趣