使用.pkl文件中的export_graphviz可视化模型

时间:2017-08-14 14:51:10

标签: machine-learning scikit-learn random-forest

我已将模型导出到.pkl文件中。现在我尝试通过Joblib导入sklearn.model_selection._search.GridSearchCV导入sklearn.tree

但是我无法使用export_graphvizexport_graphviz导入tree_,因为它希望import boto3 session = boto3.Session( region_name='eu-west-1', profile_name='myprofile' ) ec2 = session.client('ec2') response = ec2.describe_instances() obj_number = len(response['Reservations']) for objects in xrange(obj_number): try: z = response['Reservations'][objects]['Instances'][0]['Tags'][0]['Key'] except KeyError as e: untagged_instanceid = response['Reservations'][objects]['Instances'][0]['InstanceId'] untagged_state = response['Reservations'][objects]['Instances'][0]['State']['Name'] print("InstanceID: {0}, RunningState: {1}".format(untagged_instanceid, untagged_state)) 作为第一个参数。

有没有办法做到这一点?这是我的代码:

  

export_graphviz(型号,out_file = “out.dot”)   Traceback(最近一次调用最后一次):     文件“”,第1行,in     在export_graphviz中输入文件“/home/anaconda3/lib/python3.6/site-packages/sklearn/tree/export.py”,第433行       递归(decision_tree.tree_,0,criterion = decision_tree.criterion)   AttributeError:'GridSearchCV'对象没有属性'tree _'

1 个答案:

答案 0 :(得分:0)

你是如何定义gridSearchCV并训练它的?

很可能通过以下方式访问gridSearchCV中的基础树: -

decision_tree.best_estimator_.tree_ 

如果decision_tree是gridsearchcv对象

相关问题