R: 对由 caret 包构建的逻辑模型进行 car::Anova type II 和 III 测试

Question

我正在构建一个广义逻辑回归模型，如下所示:

library(mlbench)
data(PimaIndiansDiabetes)

library(caret)
trControl <- trainControl(method = "repeatedcv",
                          repeats = 3,
                          classProbs = TRUE,
                          number = 10, 
                          savePredictions = TRUE,
                          summaryFunction = twoClassSummary)

caret_model <- train(diabetes~., 
                     data=PimaIndiansDiabetes, 
                     method="glm", 
                     trControl=trControl)

那我要表演 Anova II 类和 III 类测试为:

library(car)
car::Anova(caret_model , type=2)

我收到以下错误:

Error in UseMethod("vcov") : no applicable method for 'vcov' applied to an object of class "c('train', 'train.formula')

但是，如果我使用函数 glm改为构建模型，就可以了:

glm_fit <- glm(diabetes~., data=PimaIndiansDiabetes, family=binomial)
car::Anova(glm_fit, type=2)

那么，我该如何执行 Anova我的 II 型和 III 型测试 caret模型？

Answer 1

类“火车”没有方法方差分析，所以要做你需要的:

car::Anova(caret_model$finalModel, type=2)
Analysis of Deviance Table (Type II tests)

Response: .outcome
         LR Chisq Df Pr(>Chisq)    
pregnant   15.233  1  9.505e-05 ***
glucose   114.927  1  < 2.2e-16 ***
pressure    6.548  1   0.010502 *  
triceps     0.008  1   0.928500    
insulin     1.742  1   0.186918    
mass       40.779  1  1.704e-10 ***
pedigree   10.340  1   0.001302 ** 
age         2.522  1   0.112253

这类似于:

car::Anova(glm_fit, type=2)
Analysis of Deviance Table (Type II tests)

Response: diabetes
         LR Chisq Df Pr(>Chisq)    
pregnant   15.233  1  9.505e-05 ***
glucose   114.927  1  < 2.2e-16 ***
pressure    6.548  1   0.010502 *  
triceps     0.008  1   0.928500    
insulin     1.742  1   0.186918    
mass       40.779  1  1.704e-10 ***
pedigree   10.340  1   0.001302 ** 
age         2.522  1   0.112253