获取梯度提升树GDBT模型的重要特征

获取梯度提升树GDBT模型的重要特征

 模型训练之后，想要得到比较重要的特征，可以通过python的sklearn包来实现。

python实现代码如下所示：

 GradientBoostingClassifier.py

# -*- coding: utf-8 -*-
import IOUtil as iou
import numpy as np
import pandas as pd
from sklearn import svm
from sklearn import metrics
import csv
 
def report():
    
    report = metrics.classification_report(y_test, y_predict)
    m = metrics.confusion_matrix(y_test, y_predict)
    print (report)
    print (m)
      
 
def result():
 
    pred= model.predict_proba(X_test)
    result= pd. DataFrame(pred)
    result.to_csv('result_GradientBoostingClassifier.csv')
    
#### 训练
print ("GradientBoostingClassifier Start")
 
X = iou.readArray("X_train.csv")
y = iou.readArray("y_train.csv")
y = y.ravel()
print ("GradientBoostingClassifier Data OK")
 
from sklearn.ensemble import GradientBoostingClassifier
model = GradientBoostingClassifier()
 
model.fit(X, y)
 
n=model.feature_importances_#模型的重要特征
print(n)
 
 
###### 预测
 
X_test = iou.readArray("X_test.csv")
y_test = iou.readArray("y_test.csv")
 
print ("GradientBoostingClassifier predict...")
y_predict = model.predict(X_test)
 
 
result()
report()
 
 
print ("GradientBoostingClassifier Done!!!!!!")

 

 读取文件的代码如下所示：

IOUtil.py 

import csv
import numpy as np
from sklearn.externals import joblib
 
 
def readArray( file ):
    Data = []
    reader = csv.reader(open(file),delimiter=',', quotechar='\'')
 
    for row in reader:
        row = [float(x) for x in row]
        Data.append(row)
 
    Data = np.array(Data)
 
    return Data