向 igraph 的决策树添加多个额外标签

Question

假设决策树通过 R 包可视化 igraph。

library(igraph)
n_of_vertices = 31
g = graph.tree(n_of_vertices, 2)
edge_labels = rep(c("Yes","No"), times=n_of_vertices/2)
plot(g,
    layout = layout.reingold.tilford(g, root=1),
    asp=0.5,
    vertex.shape="square",
    vertex.label=NA,
    vertex.size=6,
    vertex.color="black",
    vertex.frame.color="white",
    edge.width=3,
    edge.arrow.mode=0,
    edge.label=edge_labels,
    edge.label.family="sans",
    edge.label.color="black",
    edge.label.cex=0.75)

我正在寻找一种方法来在每个顶点行的左侧添加额外的标签，例如 "Decision 1"、"Decision 2" 等。我将这些标签以红色字体添加到下面的例子。

Answer 1

magickr 可用于操纵绘图并用文本进行注释。

library(igraph)
library(magick)

n_of_vertices = 31
g = graph.tree(n_of_vertices, 2)
edge_labels = rep(c("Yes","No"), times=n_of_vertices/2)

fig <- image_graph(width = 600, height = 600, res = 96)
#img <- image_draw(frink)

plot(g,
     layout = layout.reingold.tilford(g, root=1),
     asp=0.5,
     vertex.shape="square",
     vertex.label=NA,
     vertex.size=6,
     vertex.color="black",
     vertex.frame.color="white",
     edge.width=3,
     edge.arrow.mode=0,
     edge.label=edge_labels,
     edge.label.family="sans",
     edge.label.color="black",
     edge.label.cex=0.75)

dev.off()

out <- fig %>% 
  image_annotate("Decision Tree", size = 20, color = "black",
                              degrees = 0, location = "+250+100") 


print(out)

Answer 2

您可以只使用text功能。

## Your tree plot
plot(g,
    layout = layout.reingold.tilford(g, root=1),
    asp=0.5,
    vertex.shape="square",
    vertex.label=NA,
    vertex.size=6,
    vertex.color="black",
    vertex.frame.color="white",
    edge.width=3,
    edge.arrow.mode=0,
    edge.label=edge_labels,
    edge.label.family="sans",
    edge.label.color="black",
    edge.label.cex=0.75)

text(x=-1.2, y=c(0.5, 0, -0.5), 
    labels=c("Decision 1", "Decision 2", "Decision 3"),
    col="red")

这里的技巧之一是决定在哪里放置文本。我如何获得 x,y 坐标来放置标签？我用了locator。这使您只需单击图表上的一个点并找到它的坐标。您可能需要稍微调整或使用 cex 调整字体大小。

向 igraph 的决策树添加多个额外标签

Adding multiple extra labels to decision tree of igraph

label

r

decision-tree

igraph