请选择 进入手机版 | 继续访问电脑版
楼主: 插花好玩
3230 4

[SAS EM] 在EM中如何评判决策树模型的优劣 [推广有奖]

  • 0关注
  • 0粉丝

高中生

52%

还不是VIP/贵宾

-

威望
0
论坛币
852 个
通用积分
0
学术水平
0 点
热心指数
0 点
信用等级
0 点
经验
162 点
帖子
6
精华
0
在线时间
54 小时
注册时间
2016-12-24
最后登录
2020-5-16

插花好玩 发表于 2018-7-17 23:22:28 |显示全部楼层 |坛友微信交流群
相似文件 换一批

+2 论坛币
k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群
赵安豆老师微信:zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

感谢您参与论坛问题回答

经管之家送您两个论坛币!

+2 论坛币
在EM13.1中运行一遍系统自带数据HMEQ(信贷风险模型),使用参数如下:
1)        Date Partition: Training -70; Validation -30; Test -0.
2)        Decision Tree: Interval Target Criterion-ProbF; Nominal Target Criterion -ProbChisq; Ordinal Target Criterion-Entropy; Maximum Depth-6; Leaf size-5.
结果如图。

我的问题:
1)        如何评判决策树模型的优劣?
2)        在哪里查看模型准确性和置信区间?
二维码

扫码加我 拉你入群

请注明:姓名-公司-职位

以便审核进群资格,未注明则拒绝

关键词:决策树 树模型 Validation CRITERION partition

SAS EM-DT.PNG
插花好玩 发表于 2018-7-18 07:18:08 |显示全部楼层 |坛友微信交流群
结果说明:Score Rankings Matrix — The Score Rankings Matrix plot overlays the selected statistics for standard, baseline and best models in a lattice that is defined by the training and validation data sets. The Score Ranking Matrix plots
Cumulative Lift
Lift
Gain
% Response
Cumulative % Response
% Captured Response
Cumulative % Captured Response.

使用道具

插花好玩 发表于 2018-7-18 07:20:48 |显示全部楼层 |坛友微信交流群
结果说明2   Leaf Statistics — displays a bar chart of summary statistics for the leaves of the currently selected subtree. For binary targets, bars represent the observed probability of event for each terminal leaf in the tree. If there is validation data, then there will be a second bar for each node that represents the probability of event in the validation data.

使用道具

插花好玩 发表于 2018-7-18 07:22:33 |显示全部楼层 |坛友微信交流群
结果说明3   Tree Map — displays a compact graphical display of the tree.

A default tree in the Tree Map plot has the following properties:
The nodes are colored according to the proportion of observations in the training data for that node that match the target value.
The node width is proportional to the number of observations in the node.
If you select a node in the Tree Map window, then the corresponding node is selected in the Tree window. If a leaf node is selected, then the corresponding bars in the Leaf Statistics plot are selected.

使用道具

插花好玩 发表于 2018-7-18 07:27:39 |显示全部楼层 |坛友微信交流群
结果说明4

Fit Statistics Table

  
The Fit Statistics table displays the following statistics for the training, validation, and test data sets (if available):
_NOBS_ — Sum of Frequencies
_SUMW_ — Sum of Case Weights Times Freq
_MISC_ — Misclassification Rate
_MAX_ — Maximum Absolute Error
_SSE_ — Sum of Square Errors
_ASE_ — Average Sum of Squares
_RASE_ — Root Average Sum of Squares
_DIV_ — Divisor for ASE
_DFT_ — Total Degrees of Freedom
_APROF_ — Average Profit for Target
_PROF_ — Total Profit for Target
_PASE_ — Average Squared Error with Priors
_PMISC_ — Misclassification Rate with Priors

使用道具

您需要登录后才可以回帖 登录 | 我要注册

本版微信群
加好友,备注cda
拉您进交流群

京ICP备16021002-2号 京B2-20170662号 京公网安备 11010802022788号 论坛法律顾问:王进律师 知识产权保护声明   免责及隐私声明

GMT+8, 2024-3-28 18:45