R使用caret :: resample进行模型无法正常工作和bwplot错误

Question

我无法在不同的分类模型上获得重新采样的结果

在bwplot中我总是有这个错误

  

％模型中的模型％出错：对象＆＃39;模型＆＃39;找不到

当我训练模型时，一切都很好

重新采样仅返回：

  

致电：resamples.default（x = lst_model）

模型：ADA，RDF，MAR，SVM，NNET 重新采样数量：10 绩效指标：
时间估计：一切，最终模型适合 这是我的代码

            set.seed(7777)
            cctrlg <- trainControl(method = "cv", number = 10, returnResamp = "final",

                                   classProbs = TRUE, summaryFunction = twoClassSummary)

            model_glm_cv= train(var.imp.formula,data=data.train, method ="glm", family= binomial,preProc = c("center", "scale"),trControl=cctrlg)

            #Modele MARS
            set.seed(7777)
            cctrle <- trainControl(method = "cv", number = 10, returnResamp = "final", classProbs = TRUE)

            marsGrid <-  expand.grid(nprune = c(2, 4, 6, 8, 10),   degree = c(1, 2, 3))

            model_mars3_cv = train( var.imp.formula, data=data.train, method='earth',  tuneGrid = marsGrid, trControl =cctrle,preProc = c("center", 
            #Modele random forest
            set.seed(7777)
            seeds <- vector(mode = "list", length = nrow(data.train) + 1)

            seeds <- lapply(seeds, function(x) 1:20)

            cctrl1 <- trainControl(method = "cv", number = 10, returnResamp = "final",

                                   classProbs = TRUE, 
                                   summaryFunction = twoClassSummary,

                                    seeds = seeds)
            model_rf_cv = train(var.imp.formula, data=data.train, method='rf', ntree=400, tuneLength=5, trControl = cctrl1, preProc = c("center", 
            #Modele ADAboost M1
            set.seed(7777)
            gridAda <- expand.grid(mfinal = (1:3)*3, maxdepth = c(1, 3),coeflearn = c("Breiman", "Freund", "Zhu"))

            cctrlada <- trainControl(method = "cv", number = 10, returnResamp = "final",  classProbs = TRUE, 

                                   summaryFunction = twoClassSummary, seeds = seeds)

            model_ada_cv= train(var.imp.formula, data=data.train, method='AdaBoost.M1', tuneLength=2 ,tuneGrid = gridAda,trControl = cctrlada, preProc 
         "scale"))

            #Modele XGB
            set.seed(7777)

            xgbGrid <- expand.grid(nrounds = c(1, 10), 

                                   max_depth = 2,
                                   eta = 0.30,
                                   rate_drop = 0.10,
                                   skip_drop = 0.10,
                                   colsample_bytree = 0.90,
                                   min_child_weight = 2,
                                   subsample = 0.75,
            gamma = 0.10)
            cctrlx <- trainControl(method = "cv", number = 10, returnResamp = "final", classProbs = TRUE, summaryFunction = twoClassSummary)

            model_xgb_cv= train(var.imp.formula, data=data.train, method='xgbDART', tuneLength=5, tuneGrid=xgbGrid,trControl = cctrlx, verbose=F, 
            "scale"))



            #modele nnet
            set.seed(7777)
            cctrlnnet <- trainControl(method = "cv", number = 10, returnResamp = "final", classProbs = T)

            model_nnet_cv <- train( var.imp.formula, data=data.train, method = "nnet", trControl = cctrlnnet,

                                        preProc = c("center", "scale"),trace = FALSE)

            #SVM Radiale
            set.seed(7777)
            cctrlsv <- trainControl(method = "cv", number = 10, returnResamp = "final", classProbs = T)

            model_svmRadial_cv = train(var.imp.formula, data=data.train, method='svmRadial', tuneLength=15, tuneGrid = data.frame(.C = c(.25, .5, 1),

            .sigma = .05), trControl = cctrlsv, preProc = c("center", "scale"))

            #"prepare liste de comparaison et les compare avec resamples

            lst_model<-list(ADA=model_ada_cv, RDF=model_rf_cv, 

                            MAR=model_mars3_cv, SVM=model_svmRadial_cv,

                            NNET=model_nnet_cv)#,XGB=model_xgb_cv,        RGL=model_glm_cv)

            lst_model
            models_compare <- caret::resamples(lst_model)

            sum_model<-summary(models_compare)

            sum_model

            # bwplots de comparaison des modeles

            scales <- list(x=list(relation="free"), y=list(relation="free"))

            bwplot(models_compare, scales=scales)