基于bootstrap的预测概率与观测概率的p值比较

给定样本数据 dat 下面,我将感谢您的帮助:

(1)检查下面的方法 vignette 从logistic回归计算基于bootstrap的预测是正确的,如果我的方法有任何错误,可以帮助纠正。

(2)计算 基于bootstrap的p值比较观测概率和预测概率。

#我的示例数据:

ldose <- rep(0:5, 2)
numdead <- c(1, 4, 9, 13, 18, 20, 0, 2, 6, 10, 12, 16)
sex <- factor(rep(c("M", "F"), c(6, 6)))
SF <- cbind(numdead, numalive = 20-numdead)
dat<-data.frame(ldose, numdead, sex, SF)
tibble::rowid_to_column(dat, "indices")

new.data <- data.frame(ldose = 20, sex = "F")

#执行引导:

#Here I would appreciate any correction if something is not correct in my approach

temp.out<-function(dat, indices, new.data) {
     d<-dat[indices, ] 
     fit1<- glm(SF ~ sex*ldose, family = binomial (link = logit), data = d)  
     return(predict(fit1, new.data, type="response"))
}

results <- boot::boot(dat, temp.out, 1000, sim = "permutation")

boot::boot.ci(results, conf = 0.95, type = "all") #this fails

     Error in model.frame.default(formula = SF ~ sex * ldose, data = d, drop.unused.levels = TRUE) : 
      variable lengths differ (found for 'sex')

boot::boot.ci(results, conf = c(0.90, 0.95), type = c("perc")) #this works

#计算基于bootstrap的p值,比较观测概率(例如0.45)和预测概率(基于bootstrap算法):

#Here I would appreciate any help to calculate the p-value

提前谢谢你的帮助。如果有什么不清楚的地方,请告诉我。