R 错误：期望单个字符串值：[type=character;范围=5]

Question

我正在使用 R 编程语言。我在 R 中使用“CORELS”库，这里是一个例子（CORELS 是一个类似于决策树的统计模型）：

library(corels)

logdir <- tempdir()
rules_file <- system.file("sample_data", "compas_train.out", package="corels")
labels_file <- system.file("sample_data", "compas_train.label", package="corels")
meta_file <- system.file("sample_data", "compas_train.minor", package="corels")

stopifnot(file.exists(rules_file),
          file.exists(labels_file),
          file.exists(meta_file),
          dir.exists(logdir))

corels(rules_file, labels_file, logdir, meta_file,
       verbosity_policy = "silent",
       regularization = 0.015,
       curiosity_policy = 2,   # by lower bound
       map_type = 1)       # permutation map
cat("See ", logdir, " for result file.")

输出可以在这里查看：

OPTIMAL RULE LIST
if ({sex:Male,juvenile-crimes:>0}) then ({recidivate-within-two-years:Yes})
else if ({priors:>3}) then ({recidivate-within-two-years:Yes})
else ({recidivate-within-two-years:No})

我的问题：对于上述函数的语法如何工作，我仍然有些困惑。例如，我正在尝试在“iris”数据集上使用上述函数：

data(iris)
head(iris)

  Sepal.Length Sepal.Width Petal.Length Petal.Width Species
1          5.1         3.5          1.4         0.2  setosa
2          4.9         3.0          1.4         0.2  setosa
3          4.7         3.2          1.3         0.2  setosa
4          4.6         3.1          1.5         0.2  setosa
5          5.0         3.6          1.4         0.2  setosa
6          5.4         3.9          1.7         0.4  setosa

现在，我尝试在 iris 数据集上应用“corles”函数：

logdir <- tempdir()
rules_file <- iris
labels_file <- colnames(iris)

corels(rules_file, labels_file, logdir, 
       verbosity_policy = "silent",
       regularization = 0.015,
       curiosity_policy = 2,   # by lower bound
       map_type = 1)       # permutation map
cat("See ", logdir, " for result file.")

但这会产生以下错误：

Error in corels(rules_file, labels_file, logdir, verbosity_policy = "silent",  : 
  Expecting a single string value: [type=character; extent=5].

谁能告诉我如何解决这个错误？

谢谢

参考资料：

• https://www.rdocumentation.org/packages/corels/versions/0.0.2/topics/corels
• https://cran.r-project.org/web/packages/corels/index.html

Answer 1

问题是您试图将数据集传递给corels() 函数，但它希望您传递文件名。 corels 软件的 github 的 Data Format 部分详细介绍了如何格式化数据结构。这些应该保存在文件中，然后传递给corels()。

数据文件中的每一行都应该包含一个用大括号括起来的规则名称，{ 和 }，后跟一个 0 或 1 以指示该规则是否适用于给定的数据行.

这里有一个简单的脚本来构造一些：

library(corels)

# Write a function to take a dataset & expression and format it correctly
make_rule <- function(data,expr){
  rule_name <- deparse1(substitute(expr))
  rule_name <- gsub(" ","",rule_name)
  out <- eval(substitute(expr),data)
  paste0("{",rule_name,"} ",paste0(1*out,collapse=" "))
}

# Create names for our rule/label files

rules_file <- "rule_data"
labels_file <- "label_data"

# Create some example rules (must always be binary operations)
iris_rules <- c(
  make_rule(iris,Sepal.Length < 5.84),
  make_rule(iris,Sepal.Width < 3.05),
  make_rule(iris,Petal.Length < 3.76),
  make_rule(iris,Petal.Width < 1.2)
)

#Label data appropriately. Must be a pair of rules
# where the first is the negative option & the 2nd is the
# positive outcome. Here we want to know how to find when
# the flower is a setosa
iris_labels <- c(
  make_rule(iris,Species != "setosa"),
  make_rule(iris,Species == "setosa")
)

# Save the data in the files
writeLines(iris_rules,rules_file)
writeLines(iris_labels,labels_file)


# reference the files and set verbosity to high so
# we get the full output
corels(rules_file, labels_file, ".",
       verbosity_policy = "loud")