SAS做逻辑回归时分类变量的分类标准

是 否 +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~ !

立即领取

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

SAS做逻辑回归时对分类变量进行分类有如下两个标准：   
       WOE标准
       GINI RATIO标准
这两个标准我在网上怎么查不到资料呢？哪位大侠给点建议？怎么可以尽快了解这两个标准呢？非常感谢！

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

分享0 收藏0 回帖

关键词：分类变量 逻辑回归 ratio 非常感谢 woe 网上 资料

Weight of Evidence (WOE) measures the relative risk of an attribute. The value depends on the value of the binary target variable, which is either event or nonevent.

The predictive power of a characteristic, its ability to separate high-risk applicants from low-risk ones, is assessed by its Information Value or Gini Score. The Information Value is the weighted sum of the WOE of the characteristic's attributes. The weight is the difference between the proportion of nonevents and the proportion of events in the respective attribute.

The Gini Score is computed as follows:

1.Sort the data by the descending order of the proportion of events in an attribute. Suppose a characteristic has m attributes. Then, the sorted attributes are expressed as groups 1, 2, ...m. Each group corresponds to an attribute and group 1 has the highest proportion of events.

2.For each of these sorted groups, calculate the number of events and nonevents in group i.

哦，名白了啦

davil2000 发表于 2012-11-7 10:59
The Gini Score is computed as follows:

1.Sort the data by the descending order of the proportion  ...

非常感谢你的回答！我还想再了解一些...
那WOE究竟是一个什么样的标准？
Gini Score是the proportion of events in an attribute？那这个分数都是小数吧。
Gini Ratio等于什么呢？

另外，想问一下您，上图中红框里面的数值是怎么计算出来的呢？谢谢...

图在这里，不好意思...
这个是我对一个变量分类之后产生的统计数据。红框里面的指标不知道是怎么计算出来的。