plot(
spe.ch.beta2$height,
nrow(spe):2,
type = "S",
main = "Fusion levels - Chord - Beta-flexible",
ylab = "k (number of clusters)",
xlab = "h (node height)",
col = "grey"
)
text(spe.ch.beta2$height,
nrow(spe):2,
nrow(spe):2,
col = "red",
cex = 0.8)
text(spe.ch.UPGMA$height,
nrow(spe):2,
nrow(spe):2,
col = "red",
cex = 0.8)
# Plot the fusion level values of the Ward clustering
plot(
spe.ch.ward$height,
nrow(spe):2,
type = "S",
main = "Fusion levels - Chord - Ward",
ylab = "k (number of clusters)",
xlab = "h (node height)",
col = "grey"
)
text(spe.ch.ward$height,
nrow(spe):2,
nrow(spe):2,
col = "red",
cex = 0.8)
# Plot the fusion level values of the beta-flexible clustering
# (beta = -0.25)
Hint Observe how the function text() is used to print the number of groups
(clusters) directly on the graph.
What is the suggested number of groups for each method? Go back to the
dendrograms and cut them at the corresponding distances. Do the groups
obtained always make sense? Do you obtain enough groups containing a
substantial number of sites?
Remember that there is no single “truth” among these solutions. Each one may
provide some insight into the data.
76
4 Cluster Analysis