What if binary variable is very unbalanced - Statistics版 - 未名存档

本页内容为未名空间相应帖子的节选和存档，一周内的贴子最多显示50字，超过一周显示500字访问原贴

Statistics版 - What if binary variable is very unbalanced

相关主题
● 问个概念。	● 请教一个data mining 的问题： sequence data
● correlation between the explanatory variables	● 请教一个unbalanced group的问题包子答谢
● 15个包子求大牛指导做一道regression 题	● [合集] Re: 一个数据处理的问题,小女子请教牛人指教.
● An interview question	● R table问题弱问
● 求助，SAS运行proc logistic 出现离奇的现象！	● [合集] R问题求助... 谢谢
● 详细的当年walmartlabs的一道面试题。	● SAS format 的问题
● GBM in R	● 在 R 里面如何循环调用变量名
● How to deal with unbalanced data	● 请教ANOVA for binary data

相关话题的讨论汇总
话题: what话题: variable话题: unbalanced话题: gender话题: regression

进入Statistics版参与讨论

1

(共1页)

n**m 发帖数: 156	1 For example，if I have gender in the regression formula. My data has 90 Male and 10 Female. I know it is not good to have two groups so unbalanced. The estimate of gender will have relative big standard error. Is there any other downside for including it as explanatory variable? What if it is a dependent variable in logistic regression? Thanks.
c********h 发帖数: 330	2 如果是做response的话，看accuracy的时候要注意，因为认为所有都是男的，accuracy 还有0.9，一般用precision recall。另外太unbalanced的话，可能fitting不converge
y**3 发帖数: 267	3 If dependent variable, may do subsampling

1

(共1页)

进入Statistics版参与讨论

相关主题
● 请教ANOVA for binary data	● 求助，SAS运行proc logistic 出现离奇的现象！
● 请教：怎么能把Logistic regression的OR转化成probablity	● 详细的当年walmartlabs的一道面试题。
● R glmnet 大数据	● GBM in R
● 复杂分类变量recoding?	● How to deal with unbalanced data
● 问个概念。	● 请教一个data mining 的问题： sequence data
● correlation between the explanatory variables	● 请教一个unbalanced group的问题包子答谢
● 15个包子求大牛指导做一道regression 题	● [合集] Re: 一个数据处理的问题,小女子请教牛人指教.
● An interview question	● R table问题弱问

相关话题的讨论汇总
话题: what话题: variable话题: unbalanced话题: gender话题: regression

未名新帖统计// 7月16日

#	版面	帖数(主题数)
-	全站	4871 (796)
1	Military	3777 (569)
2	Stock	341 (51)
3	Joke	117 (17)
4	History	116 (3)
5	Automobile	100 (9)
6	USANews	55 (9)
7	Midlife	45 (1)
8	Headline	41 (41)
9	Dreamer	33 (13)
10	FleaMarket	32 (20)
11	Living	30 (7)

* 这里只显示发帖超过25的版面，努力灌水吧:-)