《应用预测建模》Max Kuhn P300页代码运行问题

1. ##我们写了两个封装函数##对于准确度、Kappa值、ROC曲线下面积、敏感度和特异度：fiveStats <- function(...) c(twoClassSummary(...),
   
2.                              defaultSummary(...))
   
3. ## 不包含ROC曲线下面积：
   
4. fourStats <- function (data, lev = levels(data$obs), model = NULL)
   
5. {
   
6.      accKapp <- postResample(data[, "pred"], data[, "obs"])
   
7.      out <- c(accKapp,
   
8.               sensitivity(data[, "pred"], data[, "obs"], lev[1]),
   
9.               specificity(data[, "pred"], data[, "obs"], lev[2]))
   
10.      names(out)[3:4] <- c("Sens", "Spec")
    
11.      out
    
12. }
    
13. ctrl <- trainControl(method = "cv",
    
14.                      classProbs = TRUE,
    
15.                      summaryFunction = fiveStats,
    
16.                      verboseIter = TRUE)
    
17. ctrlNoProb <- ctrl
    
18. ctrlNoProb$summaryFunction <- fourStats
    
19. ctrlNoProb$classProbs <- FALSE
    
20. set.seed(1410)
    
21. rfFit <- train(CARAVAN ~ ., data = trainingInd,
    
22.                method = "rf",
    
23.                trControl = ctrl,
    
24.                ntree = 1500,
    
25.                tuneLength = 5,
    
26.                metric = "ROC")+ Fold01: mtry=  2
    
27. - Fold01: mtry=  2
    
28. + Fold01: mtry=  8
    
29. - Fold01: mtry=  8
    
30. + Fold01: mtry= 34
    
31. - Fold01: mtry= 34
    
32. + Fold01: mtry=145
    
33. - Fold01: mtry=145
    
34. + Fold01: mtry=606
    
35. - Fold01: mtry=606
    
36. + Fold02: mtry=  2
    
37. - Fold02: mtry=  2
    
38. + Fold02: mtry=  8
    
39. - Fold02: mtry=  8
    
40. + Fold02: mtry= 34
    
41. - Fold02: mtry= 34
    
42. + Fold02: mtry=145
    
43. - Fold02: mtry=145
    
44. + Fold02: mtry=606
    
45. - Fold02: mtry=606
    
46. + Fold03: mtry=  2
    
47. - Fold03: mtry=  2
    
48. + Fold03: mtry=  8
    
49. - Fold03: mtry=  8
    
50. + Fold03: mtry= 34
    
51. - Fold03: mtry= 34
    
52. + Fold03: mtry=145
    
53. - Fold03: mtry=145
    
54. + Fold03: mtry=606
    
55. - Fold03: mtry=606
    
56. + Fold04: mtry=  2
    
57. - Fold04: mtry=  2
    
58. + Fold04: mtry=  8
    
59. - Fold04: mtry=  8
    
60. + Fold04: mtry= 34
    
61. - Fold04: mtry= 34
    
62. + Fold04: mtry=145
    
63. - Fold04: mtry=145
    
64. + Fold04: mtry=606
    
65. - Fold04: mtry=606
    
66. + Fold05: mtry=  2
    
67. - Fold05: mtry=  2
    
68. + Fold05: mtry=  8
    
69. - Fold05: mtry=  8
    
70. + Fold05: mtry= 34
    
71. - Fold05: mtry= 34
    
72. + Fold05: mtry=145
    
73. - Fold05: mtry=145
    
74. + Fold05: mtry=606
    
75. - Fold05: mtry=606
    
76. + Fold06: mtry=  2
    
77. - Fold06: mtry=  2
    
78. + Fold06: mtry=  8
    
79. - Fold06: mtry=  8
    
80. + Fold06: mtry= 34
    
81. - Fold06: mtry= 34
    
82. + Fold06: mtry=145
    
83. - Fold06: mtry=145
    
84. + Fold06: mtry=606
    
85. - Fold06: mtry=606

复制代码

前面还有数据处理的代码，没有一起放上来了，具体可以参考p300-p302页。输出rfFit的时候，运行一直处于一个循环的状态，请教各位这是什么原因？如何解决？

你遇到下面的问题了吗？，
> rfFit <- train(CARAVAN~.,data = trainingInd,
+                method = 'rf',
+                trControl = ctrl,
+                ntree = 1500,
+                tuneLength =5,
+                metric = 'ROC')
Error: At least one of the class levels is not a valid R variable name; This will cause errors when class probabilities are generated because the variables names will be converted to  X2, X1 . Please use factor levels that can be used as valid R variable names  (see ?make.names for help).

谢谢分享