Question

我想将kernlab软件包中的KSVM用于具有加权类（或不平衡类）的R，但是该函数无法识别向量并引发.local(x, ...)错误。

我创建了一个命名向量，用于复制和粘贴数据类别的数据变量TRAIN的级别（z =“负”，“正”，“不确定”），然后按写成的方式将此向量提供给ksvm功能手册（即使没有示例进行不平衡分析）。但是，我得到了一个拼写错误。我也尝试直接在帖子（Class Weight Syntax in Kernlab?）中给出的函数中给出向量，但是给出了相同的错误。然后我试图给每堂课一个费用，这也没有用。

>

# define named vector
> VECT = c(1.0, 0.5, 0.7)
> names(VECT) = c("negative", "positive", "uncertain")
> VECT
 negative  positive uncertain
      1.0       0.5       0.7
# show levels of dataframe
> unique(TRAIN$z)
[1] negative  positive  uncertain
Levels: negative positive uncertain
# run model with named vector
> mod = ksvm(z ~ a+b, data = TRAIN,
+                type = "C-svc",
+                kernel = "rbfdot",
+                kpar = "automatic",
+                C = 1,
+                prob.model = TRUE,
+                class.weights = VECT
+              )
Error in .local(x, ...) :
  At least one level name is missing or misspelled.
# second approach
> mod = ksvm(cons ~ mr+fcn, data = TRAIN,
+                type = "C-svc",
+                kernel = "rbfdot",
+                kpar = "automatic",
+                C = 1,
+                prob.model = TRUE,
+                class.weights = c("negative" = 0.5,
+                                  "positive" = 1,
+                                  "uncertain" = 0.7)
+              )
Error in .local(x, ...) : 
  At least one level name is missing or misspelled.
# third approach
> VECT = c(1, 10, 5)
> names(VECT) = c("negative", "positive", "uncertain")
> VECT
 negative  positive uncertain 
        1        10         5 
> mod = ksvm(cons ~ mr+fcn, data = TRAIN,
+                type = "C-svc",
+                kernel = "rbfdot",
+                kpar = "automatic",
+                C = c(0.5, 1, 0.7),
+                prob.model = TRUE,
+                class.weights = VECT
+              )
Error in .local(x, ...) : 
  At least one level name is missing or misspelled.

因此，我期望有一个完整的模型，而不是Error in .local(x, ...) : At least one level name is missing or misspelled。

因为变量z只有三个值，所以我将它们粘贴到 VECT，这怎么可能是拼写错误？

如何正确地进行不平衡的SVM分析？

谢谢

Kernlab的ksvm class.weights级别名称错误

0 个答案: