最简单的方法来绘制R中两个有序列表之间排名的变化？

Question

我想知道是否有一种简单的方法来绘制R中有向二分图形式的2个列表之间元素位置的变化.例如,列表1和2是字符串的向量,不一定包含相同的字符串.内容:

list.1 <- c("a","b","c","d","e","f","g")

list.2 <- c("b","x","e","c","z","d","a")

我想生成类似于:

我在使用igraph包时有轻微的抨击,但不能轻易地构建我想要的东西,我想象并希望不应该太难.

干杯.

Answer 1

这是一个简单的功能,可以做你想要的.基本上它用于match匹配从一个向量到另一个向量的元素并arrows绘制箭头.

plotRanks <- function(a, b, labels.offset=0.1, arrow.len=0.1)
  {
  old.par <- par(mar=c(1,1,1,1))

  # Find the length of the vectors
  len.1 <- length(a)
  len.2 <- length(b)

  # Plot two columns of equidistant points
  plot(rep(1, len.1), 1:len.1, pch=20, cex=0.8, 
       xlim=c(0, 3), ylim=c(0, max(len.1, len.2)),
       axes=F, xlab="", ylab="") # Remove axes and labels
  points(rep(2, len.2), 1:len.2, pch=20, cex=0.8)

  # Put labels next to each observation
  text(rep(1-labels.offset, len.1), 1:len.1, a)
  text(rep(2+labels.offset, len.2), 1:len.2, b)

  # Now we need to map where the elements of a are in b
  # We use the match function for this job
  a.to.b <- match(a, b)

  # Now we can draw arrows from the first column to the second
  arrows(rep(1.02, len.1), 1:len.1, rep(1.98, len.2), a.to.b, 
         length=arrow.len, angle=20)
  par(old.par)
  }

一些示例图

par(mfrow=c(2,2))
plotRanks(c("a","b","c","d","e","f","g"),
          c("b","x","e","c","z","d","a"))
plotRanks(sample(LETTERS, 20), sample(LETTERS, 5))
plotRanks(c("a","b","c","d","e","f","g"), 1:10) # No matches
plotRanks(c("a", "b", "c", 1:5), c("a", "b", "c", 1:5)) # All matches
par(mfrow=c(1,1))

Answer 2

这是使用igraph函数的解决方案.

rankchange <- function(list.1, list.2){
    grp = c(rep(0,length(list.1)),rep(1,length(list.2)))
    m = match(list.1, list.2)
    m = m + length(list.1)
    pairs = cbind(1:length(list.1), m)
    pairs = pairs[!is.na(pairs[,1]),]
    pairs = pairs[!is.na(pairs[,2]),]
    g = graph.bipartite(grp, as.vector(t(pairs)), directed=TRUE)
    V(g)$color =  c("red","green")[grp+1]
    V(g)$label = c(list.1, list.2)
    V(g)$x = grp
    V(g)$y = c(length(list.1):1, length(list.2):1)
    g
}

这将构建,然后从您的向量绘制图形:

g = rankchange(list.1, list.2)
plot(g)

调整颜色方案和符号以适合使用igraph文档中详述的选项.

请注意,这未经过全面测试(仅针对您的示例数据进行了测试),但您可以看到它是如何根据代码构建二分图的.

Answer 3

使用 ggplot2：

v1 <- c("a","b","c","d","e","f","g")
v2 <- c("b","x","e","c","z","d","a")

o <- 0.05
DF <- data.frame(x = c(rep(1, length(v1)), rep(2, length(v2))),
                 x1 = c(rep(1 + o, length(v1)), rep(2 - o, length(v2))),
                 y = c(rev(seq_along(v1)), rev(seq_along(v2))),
                 g = c(v1, v2))

library(ggplot2)
library(grid)
ggplot(DF, aes(x=x, y=y, group=g, label=g)) +
  geom_path(aes(x=x1), arrow = arrow(length = unit(0.02,"npc")), 
            size=1, color="green") +
  geom_text(size=10) +
  theme_minimal() +
  theme(axis.title = element_blank(),
        axis.text = element_blank(),
        axis.ticks = element_blank(),
        panel.grid = element_blank())

这当然可以很容易地包装在一个函数中。