无法在插入符包中为 extraTrees 模型指定概率函数

Question

大家，

最近，我一直在使用 caret 包中的 extraTrees 模型。但是，我注意到使用以下脚本将 extraTrees 模型的概率函数设置为 NULL：

extratrees_para <- getModelInfo('extraTrees', regex = F)[[1]]

extratrees_para$prob

我注意到在extraTress的原始包中，它可以用来生成分类问题的概率预测。所以我想为 extratrees_para 指定 prob 函数。

extratrees_para$prob <- function(modelFit, newdata, submodels = NULL){
as.data.frame(predict(modelFit, newdata, probability = TRUE))
}
extratrees_para$type <- 'Classification'

然后我构建训练函数来构建模型

extratreesGrid <- expand.grid(.mtry=1:2,
                      .numRandomCuts=1)

  modelfit_extratrees <- train(outcome~., data=training_scaled_sel,
                         method = extratrees_para,
                         metric = "ROC",
                         trControl = trainControl(method = 'repeatedcv',
                                                  repeats=1,
                                                  classProb = T,
                                                  summaryFunction = twoClassSummary),
                         ntree = 3000, 
                         tuneGrid = extratreesGrid)

但是，我不断收到这条信息不多的错误消息

“train.default(x, y, weights = w, ...) 中的错误：无法确定最终调整参数此外：警告信息： 1: 在 nominalTrainWorkflow(x = x, y = y, wts = weights, info = trainInfo, : 重新抽样的绩效指标中存在缺失值。 2：在 train.default(x, y, weights = w, ...) 中：在聚合结果中发现缺失值

以下是我的会话信息。如果有人可以帮助我，我将不胜感激。谢谢！

sessioninfo()

R version 3.1.2 (2014-10-31)
Platform: x86_64-w64-mingw32/x64 (64-bit)

locale:
[1] LC_COLLATE=English_United States.1252 
[2] LC_CTYPE=English_United States.1252   
[3] LC_MONETARY=English_United States.1252
[4] LC_NUMERIC=C                          
[5] LC_TIME=English_United States.1252    

attached base packages:
[1] grid      stats     graphics  grDevices utils     datasets  methods  
[8] base     

other attached packages:
[1] DMwR_0.4.1           BiocInstaller_1.16.5 caret_6.0-41        
[4] ggplot2_1.0.0        lattice_0.20-29      extraTrees_1.0.5    
[7] rJava_0.9-6         

loaded via a namespace (and not attached):
[1] abind_1.4-3         bitops_1.0-6        BradleyTerry2_1.0-5
[4] brglm_0.5-9         car_2.0-24          caTools_1.17.1     
[7] class_7.3-11        codetools_0.2-9     colorspace_1.2-4   
[10] compiler_3.1.2      digest_0.6.8        e1071_1.6-4          
[13] foreach_1.4.2       gdata_2.16.1        gplots_2.17.0      
[16] gtable_0.1.2        gtools_3.4.1        iterators_1.0.7    
[19] KernSmooth_2.23-13  lme4_1.1-7          MASS_7.3-35        
[22] Matrix_1.1-4        mgcv_1.8-3          minqa_1.2.4        
[25] munsell_0.4.2       nlme_3.1-118        nloptr_1.0.4       
[28] nnet_7.3-8          parallel_3.1.2      pbkrtest_0.4-2     
[31] plyr_1.8.1          pROC_1.8            proto_0.3-10       
[34] quantmod_0.4-4      quantreg_5.11       Rcpp_0.11.4        
[37] reshape2_1.4.1      ROCR_1.0-7          rpart_4.1-8        
[40] scales_0.2.4        SparseM_1.6         splines_3.1.2      
[43] stringr_0.6.2       tools_3.1.2         TTR_0.22-0         
[46] xts_0.9-7           zoo_1.7-12

Answer 1

当我第一次添加模型时，我认为它不会生成 class 概率。我不确定为什么你的版本不起作用，但这是我添加到包中的内容：

modelInfo <- list(label = "Random Forest by Randomization",
                  library = c("extraTrees"),
                  loop = NULL,
                  type = c('Regression', 'Classification'),
                  parameters = data.frame(parameter = c('mtry', 'numRandomCuts'),
                                          class = c('numeric', 'numeric'),
                                          label = c('# Randomly Selected Predictors', '# Random Cuts')),
                  grid = function(x, y, len = NULL){
                    expand.grid(mtry = var_seq(p = ncol(x), 
                                               classification = is.factor(y), 
                                               len = len), 
                                numRandomCuts = 1:len)
                  },
                  fit = function(x, y, wts, param, lev, last, classProbs, ...) 
                    extraTrees(x, y, mtry = param$mtry, numRandomCuts = param$numRandomCuts, ...),
                  predict = function(modelFit, newdata, submodels = NULL)
                    predict(modelFit, newdata),
                  prob = function(modelFit, newdata, submodels = NULL)
                    predict(modelFit, newdata, probability = TRUE),
                  levels = function(x) x$obsLevels,
                  tags = c("Random Forest", "Ensemble Model", "Bagging", "Implicit Feature Selection"),
                  sort = function(x) x[order(x[,1]),])

无法在插入符包中为 extraTrees 模型指定概率函数

Cannot specify probability function for extraTrees model in caret package

r

machine-learning

r-caret