R中的判别分析（FDA和MDA）图

Question

我正在尝试使用 mda 和 ggplot2 包绘制灵活判别分析 (FDA) 和混合判别分析 (MDA) 的结果。我为线性判别分析（LDA）做了它，但我不知道继续。任何帮助或想法如何使用 ggplot2 对这些图进行编码？

代码：

require(MASS)
require(ggplot2)
require(mda)
require(scales)
  
irislda <- lda(Species ~ ., iris)

prop.lda = irislda$svd^2/sum(irislda$svd^2)
plda <- predict(irislda,   iris)

dataset = data.frame(species = iris[,"Species"], irislda = plda$x)
p1 <- ggplot(dataset) + geom_point(aes(irislda.LD1, irislda.LD2, colour = species, shape = species), size = 2.5) + 
  labs(x = paste("LD1 (", percent(prop.lda[1]), ")", sep=""),
       y = paste("LD2 (", percent(prop.lda[2]), ")", sep=""))
p1 

irisfda <- fda(Species ~ ., data = iris, method = mars)
 
irismda <- mda(Species ~ ., data = iris)

Answer 1

我相信这会达到您的目标。 fda 模型只有两个维度，所以它是 100% 解释的。 mda 模型有 5 个维度，所以我只展示解释最多的两个。

library(dplyr)
irisfda <- fda(Species ~ ., data = iris, method = mars)

irisfda$fit$fitted.values %>% 
  as_tibble() %>% 
  bind_cols(species = iris[,"Species"]) %>%
  ggplot() +
  geom_point(aes(V1, V2, color = species, shape = species), size = 2.5) + 
  labs(x = paste("FDA1 (", percent(irisfda$percent.explained[1]/100), ")", sep=""),
       y = paste("FDA2 (", percent(irisfda$percent.explained[2]/100 - irisfda$percent.explained[1]/100), ")", sep=""))

irismda <- mda(Species ~ ., data = iris)

irismda$fit$fitted.values %>% 
  as_tibble() %>% 
  bind_cols(species = iris[,"Species"]) %>% 
  ggplot() +
  geom_point(aes(V1, V2, color = species, shape = species), size = 2.5) + 
  labs(x = paste("MDA1 (", percent(irismda$percent.explained[1]/100), ")", sep=""),
       y = paste("MDA2 (", percent(irismda$percent.explained[2]/100 - irismda$percent.explained[1]/100), ")", sep=""))

编辑：

为了消除您看到的警告，我们可以在将矩阵传递给as_tibble 之前命名矩阵的列。此编辑不使用%>% 运算符。

colnames(irisfda$fit$fitted.values) <- c("V1", "V2")
df1 <- bind_cols(as_tibble(irisfda$fit$fitted.values),
                 species = iris[,"Species"])
ggplot(df1) +
  geom_point(aes(V1, V2, color = species, shape = species), size = 2.5) + 
  labs(x = paste("FDA1 (", percent(irisfda$percent.explained[1]/100), ")", sep=""),
       y = paste("FDA2 (", percent(irisfda$percent.explained[2]/100 - irisfda$percent.explained[1]/100), ")", sep=""))


colnames(irismda$fit$fitted.values) <- c("V1", "V2", "V3", "V4", "V5", "V6", "V7", "V8") 
df2 <- bind_cols(as_tibble(irismda$fit$fitted.values), 
                 species = iris[,"Species"])
ggplot(df2) +
  geom_point(aes(V1, V2, color = species, shape = species), size = 2.5) + 
  labs(x = paste("MDA1 (", percent(irismda$percent.explained[1]/100), ")", sep=""),
       y = paste("MDA2 (", percent(irismda$percent.explained[2]/100 - irismda$percent.explained[1]/100), ")", sep=""))

编辑 2： 看来您不想使用dplyr，所以我在此处将基本 R 函数与ggplot 绘图一起包括在内。

library(dplyr)
require(MASS)
require(ggplot2)
require(mda)
require(scales)
irisfda <- fda(Species ~ ., data = iris, method = mars)

irismda <- mda(Species ~ ., data = iris)

df1 <- cbind(data.frame(irisfda$fit$fitted.values),
                 species = iris[,"Species"])
ggplot(df1) +
  geom_point(aes(X1, X2, color = species, shape = species), size = 2.5) + 
  labs(x = paste("FDA1 (", percent(irisfda$percent.explained[1]/100), ")", sep=""),
       y = paste("FDA2 (", percent(irisfda$percent.explained[2]/100 - irisfda$percent.explained[1]/100), ")", sep=""))


df2 <- cbind(data.frame(irismda$fit$fitted.values), 
                 species = iris[,"Species"])
ggplot(df2) +
  geom_point(aes(X1, X2, color = species, shape = species), size = 2.5) + 
  labs(x = paste("MDA1 (", percent(irismda$percent.explained[1]/100), ")", sep=""),
       y = paste("MDA2 (", percent(irismda$percent.explained[2]/100 - irismda$percent.explained[1]/100), ")", sep=""))