O'Reilly logo

Practical Predictive Analytics by Ralph Winters

Stay ahead with the world's most comprehensive technology and business learning platform.

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, tutorials, and more.

Start Free Trial

No credit card required

Cluster elbow plot

One way to determine this is by using an automated method known as the elbow method in which you plot the total within clusters sum of squares for all values within a range, and look at where an elbow method develops in the plot:

#elbow method 
set.seed(1020) 
# Compute and plot wss for k = 3 to k = 15 
df <- sapply(3:15,function(k){kmeans(y,k)$tot.withinss}) 
 
plot(3:15, df,type='b',xlab="# of clusters",ylab="Total Within Clusters SS") 

In the following plot, you might see two elbows, one at 5-6 cluster, and one at 8-9. We will assume for now that five is the optimal number, only because there is an elbow in that area, and the names given to the clusters will be easier to explain:

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, interactive tutorials, and more.

Start Free Trial

No credit card required