决策树-随机森林-两种Feature importances排序方法

第一种方法：

这种方法是唐宇迪老师视频课里的，做了很小的改动，代码如下：

# 得到特征重要性
importances = list(rf.feature_importances_)  
feature_importances = [(feature, round(importance, 2)) for feature, importance in zip(feature_list, importances)] 
print(type(feature_importances))
# 排序-降序，key参数决定按哪一列进行排序，lambda函数确定出按第二列排序
feature_importances = sorted(feature_importances, key =lambda x:x[1], reverse = True)
print(feature_importances)
# 对应进行打印
[print('Variable: {}         Importance: {}'.format(*pair)) for pair in feature_importances]

显示结果：

第二种方法：

自己写的：

importances=np.reshape(np.round(np.array(rf.feature_importances_),2),(-1,1))
feature_list=np.reshape(np.array(feature_list),(-1,1))
feature_importances=np.hstack((feature_list, importances))
#print(importances,feature_list,feature_importances)

feature_importances=pd.DataFrame(feature_importances,columns=('features','importance'))
print(feature_importances)
feature_importances.sort_values(by=['importance'],ascending = [False],inplace=True)
print(feature_importances)