将data.frame定义为距离并在R中执行分层聚类

Question

ggg <- data.frame(row.names=c("a","b","c","d","e"),var1=c("0","0","0","0","0"),var2=c("1","1","1","1","2"))

ggg_dist <- as.matrix(ggg) %>% as.dist(.)

In as.dist.default(.) : non-square matrix

class(ggg_dist)
[1] "dist"

ggg_dist
Warning message:
In df[row(df) > col(df)] <- x :
  number of items to replace is not a multiple of replacement length

 h_ggg <- hclust(ggg_dist,method="average")

Fehler in hclust(ggg_dist, method = "average") : 
  'D' must have length (N \choose 2).

我想用ggg执行分层聚类。ggg_dist是由class()中的ggg确认的距离。我想使用ggg_dist进行层次聚类，但这不起作用。它显示以上错误。我该如何解决。

我尝试过How to convert data.frame into distance matrix for hierarchical clustering?，但是在尝试调用ggg_dist时遇到相同的错误。

Answer 1

您可以使用功能dist：

ggg_dist <- dist(ggg, method = "euclidian")

结果：

ggg_dist
  a b c d
b 0      
c 0 0    
d 0 0 0  
e 1 1 1 1

Answer 2

as.dist()需要平方矩阵或data.frame。您的原始对象ggg有5行，但只有2列。

类似于以下内容将起作用。

ggg <- data.frame(row.names = c("a", "b"), 
                  var1 = c("0", "0"), 
                  var2 = c("1", "1"))

ggg_dist <- as.dist(ggg)

h_ggg <- hclust(ggg_dist, method="average")
h_ggg
#> 
#> Call:
#> hclust(d = ggg_dist, method = "average")
#> 
#> Cluster method   : average 
#> Number of objects: 2

^{由reprex package（v0.3.0）在2020-05-27创建}