特征重要性

#检测重要特征rf = RandomForestClassifier()

rf.fit(X, y)

f, ax = plt.subplots(figsize=(7, 5))

ax.bar(range(len(rf.feature_importances_)),rf.feature_importances_)

ax.set_title("Feature Importances")

f.show()

#每个例子属于哪个类的概率probs = rf.predict_proba(X)import pandas as pd

probs_df = pd.DataFrame(probs, columns=['0', '1'])

probs_df['was_correct'] = rf.predict(X) == yimport matplotlib.pyplot as plt

f, ax = plt.subplots(figsize=(7, 5))

probs_df.groupby('0').was_correct.mean().plot(kind='bar', ax=ax)

ax.set_title("Accuracy at 0 class probability")

ax.set_ylabel("% Correct")

ax.set_xlabel("% trees for 0")

f.show()

特征重要性

forest=RandomForestClassifier(n_estimators=10,n_jobs=-1,random_state=9)

forest.fit(x_train,y_train)

importances=forest.feature_importances_

print('每个维度对应的重要性因子：\n',importances)

indices = np.argsort(importances)[::-1]# a[::-1]让a逆序输出print('得到按维度重要性因子排序的维度的序号：\n',indices)

most_import = indices[:3]#取最总要的3个print(x_train[:,most_import])

特征重要性

from sklearn.datasets import load_boston

from sklearn.ensemble import RandomForestRegressor

import numpy as np

#Load boston housing dataset as an example

boston = load_boston()

X = boston["data"]

print type(X),X.shape

Y = boston["target"]

names = boston["feature_names"]

print names

rf = RandomForestRegressor()

rf.fit(X, Y)

print "Features sorted by their score:"

print sorted(zip(map(lambda x: round(x, 4), rf.feature_importances_), names), reverse=True)

参数调优

param_test1= {'n_estimators':list(range(10,71,10))}    #对参数'n_estimators'进行网格调参

gsearch1= GridSearchCV(estimator = RandomForestClassifier(min_samples_split=100,min_samples_leaf=20,max_depth=8,max_features='sqrt' ,random_state=10), param_grid =param_test1, scoring='roc_auc',cv=5) 

gsearch1.fit(X,y) 

gsearch1.grid_scores_,gsearch1.best_params_, gsearch1.best_score_    #输出调参结果，并返回最优下的参数

#输出结果如下：

([mean:0.80681, std: 0.02236, params: {'n_estimators': 10}, 

  mean: 0.81600, std: 0.03275, params:{'n_estimators': 20}, 

  mean: 0.81818, std: 0.03136, params:{'n_estimators': 30}, 

  mean: 0.81838, std: 0.03118, params:{'n_estimators': 40}, 

  mean: 0.82034, std: 0.03001, params:{'n_estimators': 50}, 

  mean: 0.82113, std: 0.02966, params:{'n_estimators': 60}, 

  mean: 0.81992, std: 0.02836, params:{'n_estimators': 70}], 

{'n_estimators':60},  0.8211334476626017)

#多个特征的网格搜索，如下所示

param_test2= {'max_depth':list(range(3,14,2)),'min_samples_split':list(range(50,201,20))} 

gsearch2= GridSearchCV(estimator = RandomForestClassifier(n_estimators= 60, min_samples_leaf=20,max_features='sqrt' ,oob_score=True,random_state=10),  param_grid = param_test2,scoring='roc_auc',iid=False, cv=5) 

gsearch2.fit(X,y) 

gsearch2.grid_scores_,gsearch2.best_params_, gsearch2.best_score_ 

#通过查看袋外准确率（oob）来判别参数调整前后准确率的变化情况

rf1= RandomForestClassifier(n_estimators= 60, max_depth=13, min_samples_split=110,  min_samples_leaf=20,max_features='sqrt' ,oob_score=True,random_state=10) 

rf1.fit(X,y) 

print(rf1.oob_score_)   

#通过每次对1-3个特征进行网格搜索，重复此过程直到遍历每个特征，并得到最终的调参结果。