ggplot2混淆矩阵geom_text标签

Question

我已经使用ggplot和geom_text标签的比例在R中绘制了一个混淆矩阵（预测5个结果）。

geom_text(aes(label = percent(Freq/sum(Freq)))用代码编写的方式，它显示每个框的频率除以所有观测值的和，但是我要做的是获取每个框的频率除以每个参考的和频率。 >

换句话说，不是A，A = 15.8％，
应该是A，A = 15.8％/（0.0％+ 0.0％+ 0.0％+ 0.0％+ 15.8 %%）= 100.0％

library(ggplot2)
library(scales)

valid_actual <- as.factor(c("A","B","B","C","C","C","E","E","D","D","A","A","A","E","E","D","D","C","B"))
valid_pred   <- as.factor(c("A","B","C","C","E","C","E","E","D","B","A","B","A","E","D","E","D","C","B"))

cfm <- confusionMatrix(valid_actual, valid_pred)


ggplotConfusionMatrix <- function(m){
  mytitle <- paste("Accuracy", percent_format()(m$overall[1]),
                   "Kappa", percent_format()(m$overall[2]))
  p <-
    ggplot(data = as.data.frame(m$table) ,
           aes(x = Reference, y = Prediction)) +
    geom_tile(aes(fill = log(Freq)), colour = "white") +
    scale_fill_gradient(low = "white", high = "green") +
    geom_text(aes(x = Reference, y = Prediction, label = percent(Freq/sum(Freq)))) +
    theme(legend.position = "none") +
    ggtitle(mytitle)
  return(p)
}

ggplotConfusionMatrix(cfm)

Answer 1

问题是，据我所知，ggplot无法进行组计算。有关类似问题，请参见this recent post。

要解决您的问题，您应该利用dplyr软件包。

这应该有效

library(ggplot2)
library(scales)
library(caret)

library(dplyr)

valid_actual <- as.factor(c("A","B","B","C","C","C","E","E","D","D","A","A","A","E","E","D","D","C","B"))

valid_pred   <- as.factor(c("A","B","C","C","E","C","E","E","D","B","A","B","A","E","D","E","D","C","B"))

cfm <- confusionMatrix(valid_actual, valid_pred)

ggplotConfusionMatrix <- function(m){
  mytitle <- paste("Accuracy", percent_format()(m$overall[1]),
                   "Kappa", percent_format()(m$overall[2]))
  data_c <-  mutate(group_by(as.data.frame(m$table), Reference ), percentage = 
percent(Freq/sum(Freq)))
  p <-
    ggplot(data = data_c,
           aes(x = Reference, y = Prediction)) +
    geom_tile(aes(fill = log(Freq)), colour = "white") +
    scale_fill_gradient(low = "white", high = "green") +
    geom_text(aes(x = Reference, y = Prediction, label = percentage)) +
    theme(legend.position = "none") +
    ggtitle(mytitle)
  return(p)
}

ggplotConfusionMatrix(cfm)

结果：