R中的F_meas中错误'输入数据必须具有相同的两个级别？

Question

set.seed(1)
library(caret)
library(dslabs)
library(dplyr)

data("tissue_gene_expression")
y_<-tissue_gene_expression$y

test_index <- createDataPartition(y_, times = 1, ,p=0.5, list = FALSE)

x_train<-tissue_gene_expression$x[-test_index,]
y_train<-tissue_gene_expression$y[-test_index]

x_test<-tissue_gene_expression$x[test_index,]
y_test<-tissue_gene_expression$y[test_index]

fit<-knn3(x_train,y_train,k=1)
y_test_hat<-predict(fit,x_test,type = 'class')

F_meas(data = y_test_hat,reference = y_test)

上面是我的代码，它总是返回错误：

Error in F_meas.default(data = y_test_hat, reference = y_test, ) :
 input data must have the same two levels

尽管我已经检查了这两个数据（y_test_hat和y_test_hat）的级别，并且它们具有相同的7个级别

Answer 1

在这种情况下，两种方法是2，而不是7。因此，为了正确使用F_meas，您需要先置换小脑，结肠，子宫内膜，海马，肾脏，肝脏，胎盘，并为F_meas提供两个水平否则将无法正常工作。

#' @rdname recall
#' @importFrom stats complete.cases
#' @export
recall.default <- function(data, reference, relevant = levels(reference)[1], 
                       na.rm = TRUE, ...) {
  if (!is.factor(reference) | !is.factor(data)) 
stop("input data must be a factor")
  if (length(unique(c(levels(reference), levels(data)))) != 2) 
stop("input data must have the same two levels") # where we see two means 2
  if (na.rm) {
cc <- complete.cases(data) & complete.cases(reference)
    if (any(!cc)) {
  data <- data[cc]
  reference <- reference[cc]
}
}
  xtab <- table(data, reference)
 recall.table(xtab, relevant = relevant)
}  

希望这可以澄清。