应用函数用于glm模型列表

Question

你可以写任何一个帮我编写for循环或应用函数来运行以下代码的多个模型

模拟数据

set.seed(666)
x1 = rnorm(1000) 
x2 = rnorm(1000)
y = rbinom(1000,1,0.8)
df = data.frame(y=as.factor(y),x1=x1,x2=x2)

将数据拆分为训练和测试集

dt = sort(sample(nrow(df), nrow(df)*.5, replace = F))
trainset=df[dt,]; testset=df[-dt,]

拟合逻辑回归模型

model1=glm( y~x1,data=trainset,family="binomial")
model2=glm( y~x1+x2,data=trainset,family="binomial")

测试测试和训练中的模型准确性

我想为上面安装的多个模型循环下面提到的代码，并在每个模型的列车集和测试集中打印AUC

require(pROC)
trainpredictions <- predict(object=model1,newdata = trainset); 
trainpredictions <- as.ordered(trainpredictions)
testpredictions <- predict(object=model1,newdata = testset); 
testpredictions <- as.ordered(testpredictions)
trainauc <- roc(trainset$y, trainpredictions); 
testauc <- roc(testset$y, testpredictions)
print(trainauc$auc); print(testauc$auc)

Answer 1

只需将模型放入列表

即可

models <- list(
  model1 = glm( y~x1,data=trainset,family="binomial"),
  model2 = glm( y~x1+x2,data=trainset,family="binomial")
)

定义值提取功能

getauc <- function(model) {
  trainpredictions <- predict(object=model,newdata = trainset); 
  trainpredictions <- as.ordered(trainpredictions)
  testpredictions <- predict(object=model,newdata = testset); 
  testpredictions <- as.ordered(testpredictions)
  trainauc <- roc(trainset$y, trainpredictions); 
  testauc <- roc(testset$y, testpredictions)
  c(train=trainauc$auc, test=testauc$auc)
}

sapply()对您的列表起作用

sapply(models, getauc)
#          model1    model2
# train 0.5273818 0.5448066
# test  0.5025038 0.5146211