需要对三组人群进行itpw匹配,然后匹配后进行比较是否有差异
网上找的教程,但是跑到最后会报错,不知道该怎么改?
代码如下:
bc <-na.omit(bc)
bc$group<-factor(bc$group)
bc$race<-factor(bc$race)
bc$gender<-factor(bc$gender)
fit<-glm(group~ age + race + gender,
family = binomial("logit"),
data = bc)
summary(fit)
exp(confint(fit))
exp(coef(fit))
library(ipw)
w1 <-ipwpoint(
exposure = group,
family = "multinomial",
numerator = ~ 1,
denominator = age + race + gender,
data = bc)
bc$w1<-w1$ipw.weights
fit<-glm(group~ age + race + gender,family =binomial("logit"),data = bc,weights = w1)
summary(fit1.IPTW)
exp(confint(fit1.IPTW))
exp(coef(fit1.IPTW))
运行:fit<-glm(group~ age + race + gender,family =binomial("logit"),data = bc,weights = w1)
就会警告:Warning message: In eval(family$initialize) : non-integer #successes in a binomial glm!
然后后面三行代码都没法跑了。有大佬可以教教我吗?三组的iptw网上没咋找到教程
数据保存在baidu盘里了:
链接: https://pan.baidu.com/s/19dOezf2HasxvSy8smZDWGQ 提取码: mfva 复制这段内容后打开百度网盘手机App,操作更方便哦