R geom_平铺、geom_文本和镶嵌面_包裹产生不正确的混淆矩阵

我的数据框如下所示：

method1 = c("cam","sce","cam","sce","cam","sce","cam","sce","cam","sce","cam","sce","cam","sce","cam","sce");
method2 = c("cam","cam","sce","sce","cam","cam","sce","sce","cam","cam","sce","sce","cam","cam","sce","sce");
p = c(0,0.191781,0.780822,0,0,0.082192,0.890411,0,0,0.383562,0.616438,0,0,0.054795,0.863014,0);
participant = c("LucO","LucO","LucO","LucO","ad","ad","ad","ad","arthur","arthur","arthur","arthur","evgeny","evgeny","evgeny","evgeny");
mydata = data.frame(method1,method2,p,participant);

我试图将每位参与者的困惑矩阵绘制为：

p <- ggplot( mydata, aes( x=method1, y=method2, fill=method2 ) ) + 
     geom_tile( aes( fill = mydata$p ) ) +
     geom_text( aes( fill = mydata$p, label = round( mydata$p, 3 ) * 100 ) ) +
     facet_wrap( ~participant, ncol=3, scales="free_x" );

---

p只需丢失
mydata$p
并添加
p
即可

函数的正确调用是：

library(ggplot2)
p <- ggplot( mydata, aes( x=method1, y=method2, fill=method2 ) ) + 
  geom_tile( aes( fill = p ) ) +
  geom_text( aes( fill = p, label = round( p, 3 ) * 100 ) ) +
  facet_wrap( ~participant, ncol=3, scales="free_x" );
p

> mydata
   method1 method2        p participant
1      cam     cam 0.000000        LucO
2      sce     cam 0.191781        LucO
3      cam     sce 0.780822        LucO
4      sce     sce 0.000000        LucO
5      cam     cam 0.000000          ad
6      sce     cam 0.082192          ad
7      cam     sce 0.890411          ad
8      sce     sce 0.000000          ad
9      cam     cam 0.000000      arthur
10     sce     cam 0.383562      arthur
11     cam     sce 0.616438      arthur
12     sce     sce 0.000000      arthur
13     cam     cam 0.000000      evgeny
14     sce     cam 0.054795      evgeny
15     cam     sce 0.863014      evgeny
16     sce     sce 0.000000      evgeny