0

I'am confused about how to calculate clustering performance with kmeans clustering. If in supervised learning we use confusion matrix to calculate classification performance. What about unsupervised learning, in this case i'am use kmeans? Anyone can show clustering performance with kmeans in R?

Dio Agus Nofrizal
  • 27
  • 5

1 Answers1

1

You could use several metrics like silhouette for example. I recommend you to look a little bit about other clustering performance measurements, but here you have more info about silhouette: https://scikit-learn.org/stable/auto_examples/cluster/plot_kmeans_silhouette_analysis.html?highlight=silhouette

mfalco
  • 428
  • 3
  • 14
  • If i use elbow method to select number of cluster, the score is value sum of square? – Dio Agus Nofrizal Mar 12 '21 at 10:40
  • Check this post: https://stackoverflow.com/questions/19197715/scikit-learn-k-means-elbow-criterion . But please do a little research before asking, this was the first result in google. And also accept the answer (click the tick) if your question was solved. – mfalco Mar 12 '21 at 12:29