我正在阅读Max Kuhn和Kjell Johnson的著作“应用预测建模”的例子,不幸的是,在使用train()
函数进行支持向量机交叉验证时,我卡在了其中一个GermanCredit
数据集的示例上,该数据集由caret
包提供:
library(AppliedPredictiveModeling)
library(caret)
# preparing the data
data(GermanCredit)
GermanCredit <- GermanCredit[, -nearZeroVar(GermanCredit)]
GermanCredit$CheckingAccountStatus.lt.0 <- NULL
GermanCredit$SavingsAccountBonds.lt.100 <- NULL
GermanCredit$EmploymentDuration.lt.1 <- NULL
GermanCredit$EmploymentDuration.Unemployed <- NULL
GermanCredit$Personal.Male.Married.Widowed <- NULL
GermanCredit$Property.Unknown <- NULL
GermanCredit$Housing.ForFree <- NULL
set.seed(100)
inTrain <- createDataPartition(GermanCredit$Class, p = .8)[[1]]
GermanCreditTrain <- GermanCredit[ inTrain, ]
GermanCreditTest <- GermanCredit[-inTrain, ]
# Grid selection for `sigma` and `cost` tuning parameters:
library(kernlab)
set.seed(231)
sigDist <- sigest(Class ~ ., data = GermanCreditTrain, frac = 1)
svmTuneGrid <- data.frame(.sigma = sigDist[1], .C = 2^(-2:7))
# SVM classification and cross-validation
svmFit <- train(Class ~ .,
data = GermanCreditTrain,
method = "svmRadial",
preProc = c("center", "scale"),
tuneGrid = svmTuneGrid,
trControl = trainControl(method = "repeatedcv", repeats = 5,
classProbs = TRUE))
而且它抛出了这个错误:
Error in comp(expr, env = envir, options = list(suppressUndefined = TRUE)) :
could not find function "makeCenv"
有时会出现以下错误信息:
Loading required package: class
Warning: namespace ‘compiler’ is not available and has been replaced
by .GlobalEnv when processing object ‘GermanCredit’
Error in comp(expr, env = envir, options = list(suppressUndefined = TRUE)) :
could not find function "makeCenv"
In addition: Warning message:
executing %dopar% sequentially: no parallel backend registered
然后我得知
makeCenv()
在doMC
包中,该包被建议作为并行计算或并行处理的替代方案,但是我不会选择该包,因为它在Windows平台上不可用。有其他替代方案吗?更新: 当代码在
Rstudio IDE
下运行时出现这些错误,从默认的R控制台运行时一切正常,所以问题是局限于Rstudio中的,我猜想。尽管在R控制台中时间有点长(约8分钟),我想知道如何加快速度,考虑到下面提到的硬件规格。我的sessioninfo()输出在这里(Rstudio):
R version 3.0.2 (2013-09-25)
Platform: i386-w64-mingw32/i386 (32-bit)
locale:
[1] LC_COLLATE=English_United States.1252 LC_CTYPE=English_United States.1252
[3] LC_MONETARY=English_United States.1252 LC_NUMERIC=C
[5] LC_TIME=English_United States.1252
attached base packages:
[1] datasets grid splines utils stats graphics grDevices methods
[9] base
other attached packages:
[1] proxy_0.4-10 e1071_1.6-1
[3] class_7.3-9 kernlab_0.9-19
[5] caret_5.17-7 foreach_1.4.1
[7] AppliedPredictiveModeling_1.1-4 CORElearn_0.9.42
[9] rpart_4.1-3 xtable_1.7-1
[11] knitr_1.5 texreg_1.30
[13] pastecs_1.3-15 boot_1.3-9
[15] gridExtra_0.9.1 reshape2_1.2.2
[17] plyr_1.8 scales_0.2.3
[19] ggplot2_0.9.3.1 vcdExtra_0.5-11
[21] gnm_1.0-6 vcd_1.3-1
[23] corrplot_0.73 RColorBrewer_1.0-5
[25] car_2.0-19 Hmisc_3.13-0
[27] Formula_1.1-1 cluster_1.14.4
[29] xlsx_0.5.5 xlsxjars_0.5.0
[31] rJava_0.9-5 lmPerm_1.1-2
[33] coin_1.0-23 survival_2.37-4
[35] GPArotation_2012.3-1 psych_1.3.12
[37] sos_1.3-8 brew_1.0-6
[39] data.table_1.8.10 mice_2.18
[41] nnet_7.3-7 MASS_7.3-29
[43] lattice_0.20-23
loaded via a namespace (and not attached):
[1] codetools_0.2-8 colorspace_1.2-4 dichromat_2.0-0 digest_0.6.4
[5] evaluate_0.5.1 formatR_0.10 gtable_0.1.2 iterators_1.0.6
[9] labeling_0.2 Matrix_1.1-0 modeltools_0.2-21 munsell_0.4.2
[13] mvtnorm_0.9-9996 proto_0.3-10 qvcalc_0.8-8 relimp_1.0-3
[17] stats4_3.0.2 stringr_0.6.2 tcltk_3.0.2 tools_3.0.2
默认 R 控制台中的 sessionInfo() 输出:
R version 3.0.2 (2013-09-25)
Platform: i386-w64-mingw32/i386 (32-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] datasets grDevices grid splines graphics utils stats
[8] methods base
other attached packages:
[1] e1071_1.6-1 class_7.3-9 kernlab_0.9-19 caret_5.17-7
[5] foreach_1.4.1 cluster_1.14.4 lattice_0.20-23 reshape2_1.2.2
[9] plyr_1.8 scales_0.2.3 ggplot2_0.9.3.1 lmPerm_1.1-2
[13] coin_1.0-23 survival_2.37-4 sos_1.3-8 brew_1.0-6
loaded via a namespace (and not attached):
[1] codetools_0.2-8 colorspace_1.2-4 compiler_3.0.2 dichromat_2.0-0
[5] digest_0.6.3 gtable_0.1.2 iterators_1.0.6 labeling_0.2
[9] MASS_7.3-29 modeltools_0.2-21 munsell_0.4.2 mvtnorm_0.9-9996
[13] proto_0.3-10 RColorBrewer_1.0-5 stats4_3.0.2 stringr_0.6.2
[17] tools_3.0.2
问题:
在默认的R控制台中运行良好,因此必须与
Rstudio
进行交互,基于默认R控制台和Rstudio的两个sessionInfo()输出的差异是compiler
包。奇怪的是,在CRAN中找不到这个包,我在这里找到了一个注释: http://www.inside-r.org/r-doc/compiler/compile 说load(compiler)就足够了。但是当我在Rstudio中执行时,出现了以下错误信息:Error: package ‘compiler’ was built before R 3.0.0: please re-install it
更新
最终从Rstudio中复制并粘贴编译器包库成功,但仍然需要很长时间(约8分钟),如果硬件和Windows系统适用,我会发布一个关于并行处理的单独问题,以便更早地找到答案。
- 我的笔记本电脑是2.1GHz双核处理器,3GB内存,Windows 32位,有没有想法如何使用
train()
函数进行并行处理?能否提供这方面的R代码?我将非常感激。
doMC
,我会排除这个。错误信息一直指向makeCenv()
,但这次出现了namespace ‘compiler’ is not available
,请参见上面的更新错误信息。我希望有一个解决办法。 - doctorateError: package ‘compiler’ was built before R 3.0.0: please re-install it
. - doctoratecompiler_3.0.2
版本,因为从 CRAN 复制和粘贴无法下载,但它可以正常工作。似乎这个包已经随 R 一起发布了,但由于某种原因,我在 Rstudio 的 lib 路径中有一个旧版本,导致了所有的麻烦。 - doctorate