The goal of clust431 is to do k-means clustering and hierarchical clustering.
You can install the released version of clust431 from CRAN with:
install.packages("clust431")This is a basic example which compares the original kmeans function to the one made for lab 7
library(clust431)
library(tidyverse)
#> -- Attaching packages --------------------------------------- tidyverse 1.3.0 --
#> v ggplot2 3.3.3 v purrr 0.3.4
#> v tibble 3.1.1 v dplyr 1.0.6
#> v tidyr 1.1.2 v stringr 1.4.0
#> v readr 1.4.0 v forcats 0.5.0
#> Warning: package 'tibble' was built under R version 4.0.5
#> Warning: package 'dplyr' was built under R version 4.0.5
#> -- Conflicts ------------------------------------------ tidyverse_conflicts() --
#> x dplyr::filter() masks stats::filter()
#> x dplyr::lag() masks stats::lag()
iris2 <- iris %>%
select(Sepal.Length, Sepal.Width)
k_means(iris2, 3, PCA = F)
#> $`Clustering vector`
#> [1] 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
#> [38] 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 3 2 3 2 3 2 3 3 3 3 3 3 2 3 3 3 3 3 3 3 3
#> [75] 2 2 2 2 3 3 3 3 3 3 3 3 2 3 3 3 3 3 3 3 3 3 3 3 3 3 2 3 2 2 2 2 3 2 2 2 2
#> [112] 2 2 3 3 2 2 2 2 3 2 3 2 3 2 2 3 3 2 2 2 2 2 3 3 2 2 2 3 2 2 2 3 2 2 2 3 2
#> [149] 2 3
#>
#> $`Cluster Means`
#> Sepal.Length Sepal.Width
#> 1 5.006000 3.428000
#> 2 6.812766 3.074468
#> 3 5.773585 2.692453
#>
#> $SSE
#> [1] 37.0507
#>
#> $`Total SS`
#> [1] 130.4753The chunk below is comparing the results of my function from above to the built-in function.
actual <- kmeans(iris2, 3)
actual
#> K-means clustering with 3 clusters of sizes 53, 50, 47
#>
#> Cluster means:
#> Sepal.Length Sepal.Width
#> 1 5.773585 2.692453
#> 2 5.006000 3.428000
#> 3 6.812766 3.074468
#>
#> Clustering vector:
#> [1] 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
#> [38] 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 1 3 1 3 1 3 1 1 1 1 1 1 3 1 1 1 1 1 1 1 1
#> [75] 3 3 3 3 1 1 1 1 1 1 1 1 3 1 1 1 1 1 1 1 1 1 1 1 1 1 3 1 3 3 3 3 1 3 3 3 3
#> [112] 3 3 1 1 3 3 3 3 1 3 1 3 1 3 3 1 1 3 3 3 3 3 1 1 3 3 3 1 3 3 3 1 3 3 3 1 3
#> [149] 3 1
#>
#> Within cluster sum of squares by cluster:
#> [1] 11.3000 13.1290 12.6217
#> (between_SS / total_SS = 71.6 %)
#>
#> Available components:
#>
#> [1] "cluster" "centers" "totss" "withinss" "tot.withinss"
#> [6] "betweenss" "size" "iter" "ifault"
actual$totss - actual$betweenss
#> [1] 37.0507
actual$totss
#> [1] 130.4753As we can see, we get the same SSE, clustering vectors, and cluster means. Our function gives SSE straight up, but there is more calculations needed. That being said, the built-in function has much more information.
This is a basic example which compares the actual hier_clust function to the one made for lab 7
iris3 <- iris %>%
select(-c(Species))
b = hier_clust(iris3, k=150, method = 'euclidean')
head(b, 6)
#> cluster 1 cluster 2 cluster 3 cluster 4 cluster 5 cluster 6
#> "102,143" "8,40" "1,18" "10,35" "133,129" "11,49"
a = hclust(dist(iris3))
head(a$merge, 6)
#> [,1] [,2]
#> [1,] -102 -143
#> [2,] -8 -40
#> [3,] -1 -18
#> [4,] -10 -35
#> [5,] -129 -133
#> [6,] -11 -49hier_clust(iris3, k=3, method = "manhattan")
#> cluster 1
#> "150,71,139,128,115,122,114,143,102,134,84,73,124,127,112,147,135,52,57,86,98,75,74,79,64,92,55,59,77,76,66,51,53,87,78,120,69,88,145,141,144,121,125,137,149,101,133,129,105,104,117,138,109,148,111,116,146,142,113,140"
#> cluster 2
#> "7,43,4,48,3,14,39,9,23,26,10,35,2,46,13,36,38,5,29,28,18,1,41,50,8,40,25,31,30,12,42,17,34,33,15,16,21,32,37,11,49,6,19,45,47,22,20,24,27,44"
#> cluster 3
#> "150,71,139,128,115,122,114,143,102,134,84,73,124,127,112,147,135,52,57,86,98,75,74,79,64,92,55,59,77,76,66,51,53,87,78,120,69,88,145,141,144,121,125,137,149,101,133,129,105,104,117,138,109,148,111,116,146,142,113,140,110,136,132,118,119,123,106,103,130,126,108,131"hier_clust(iris3, k=4, method = "euclidean")
#> cluster 1
#> "37,21,32,49,11,22,20,47,45,44,27,24,8,40,28,29,38,5,1,18,41,31,30,12,25,2,13,46,26,10,35,36,50,15,19,16,33,34,17,6"
#> cluster 2
#> "101,149,137,141,121,144,145,125,148,111,116,140,113,142,146,131,108,130,126,103,129,133,105,104,138,117,109"
#> cluster 3
#> "78,76,66,59,55,77,51,87,53,86,52,57,64,92,79,74,98,75,72,88,69,124,127,134,73,147,112,120,84,135,71,150,128,139,115,122,114,143,102"
#> cluster 4
#> "60,65,56,100,95,91,62,96,97,89,67,85,70,82,81,90,54,63,68,83,93,80,61,99,94,58,107"