cluster.R

install.packages("cluster")
install.packages("ggdendro")
install.packages("factoextra")
library(cluster)
library(ggdendro)
library(factoextra)

hc <- 
  df %>% 
  select(-id) %>% 
  select(-starts_with("PCB")) %>% 
  agnes(method= "ward")
hc
hc %>% ggdendrogram()

df %>% 
  select(-id) %>% 
  select(-starts_with("PCB")) %>% 
  fviz_nbclust(FUN = hcut, method = "wss")
df %>% 
  select(-id) %>% 
  select(-starts_with("PCB")) %>% 
  fviz_nbclust(FUN = hcut, method = "silhouette")
df %>% 
  select(-id) %>% 
  select(-starts_with("PCB")) %>% 
  clusGap(FUN = hcut, nstart = 25, K.max = 10, B = 50) %>% 
  fviz_gap_stat()

clusterID <- hc %>% cutree(k = 4)
table(clusterID)
hc
plot(hc, ask = FALSE, which.plots = 2)
rect.hclust(hc, k = 4, border = 2:8)

df0 <- 
  df %>% 
  mutate(clusterID = hc %>% cutree(k = 4))

hc_a <- agnes(df, method = "ward")
cutree(as.hclust(hc_a), k = 4)