r - 在 PCA 图中命名样本

Question

我有一个包含大量数据的 PCA 图，我想确定哪些样本是异常值。
当我使用

geom.ind = c("text")

然后有太多的文字，我什么也读不懂。

这是一个最小的可重现示例。 (我已经在这里使用它 tooltip with names in a PCA plot 但答案只能手动工作，而且我真的有一个很棒的数据框)

dataframe <- data_frame("c1"=c(78,89,0),"c2"=c(89,89,34),"c3"=c(56,0,4))
row.names(dataframe) <- c("name1","name2","name3")

sub <- PCA(dataframe)

pca <- fviz_pca_ind(sub, pointsize = "cos2", 
             pointshape = 21, fill = "#E7B800",
             repel = TRUE, # Avoid text overlapping (slow if many points)
             geom = c("text","point"), 
             xlab = "PC1", ylab = "PC2",label = row.names(dataframe)
             )

interactive <- ggplotly(pca,dynamicTicks = T,tooltip = c("x","y",label = list))

如您所见，我尝试使用 ggplotly() 函数来完成此操作，但这不起作用。

我想在我的图中识别样本名称 (name1,name2,name3)。我怎样才能为一个伟大的数据集做到这一点？

非常感谢你提前

Answer 1

您可以使用以下代码

library(tidyverse)
library("factoextra")
library(plotly)
library(FactoMineR)

dataframe <- data_frame("c1"=c(78,89,0),"c2"=c(89,89,34),"c3"=c(56,0,4))
row.names(dataframe) <- c("name1","name2","name3")

sub <- PCA(dataframe)

pca <- fviz_pca_ind(sub, pointsize = "cos2", 
                    pointshape = 21, fill = "#E7B800",
                    repel = TRUE, # Avoid text overlapping (slow if many points)
                    geom = c("text","point"), 
                    xlab = "PC1", ylab = "PC2",label = c("ind")
)

interactive <- ggplotly(pca,tooltip = c("x","y","colour"))

bggly <- plotly_build(interactive)
bggly$x$data[[1]]$text <- 
  with(pca$data, paste0("name: ", name, 
                        "</br></br>x: ", x, 
                        "</br>y: ", y, 
                        "</br>coord: ", coord, 
                        "</br>cos2: ", cos2, 
                        "</br>contrib: ", contrib))
bggly

在接受 this post 的帮助后作者:斯蒂芬·洛朗
对于 .csv 格式的大数据集，第一列作为行名称，您可以将其读入 df <- read.csv("Test_Data.csv", row.names = 1) ，前提是您的行名称不重复。