What is the difference between “tree-based clustering” and “tree-based classification”?

Clustering is an unsupervised learning method and classification is a supervised one.

I understand that a tree can be used for image classification since it is based on decision tree in which we have yes/no conditions. we grow a tree to classify a query image whether it belongs to a predefined group or not. But what I cannot understand is clustering using trees. In clustering, how we can adopt a decision tree to group our data in $n$ groups?

• My guess, using some kind of Cost function. Can you give examples where did you encounter it? – Royi Apr 9 '16 at 18:55
• @Drazick, since classification is supervised-based method, we have a trained and grown tree and then we apply a new piece of information and the tree tells us what class it belongs. but for clustering in which we do not have examples from which we can grow a tree, how clustering works and how we can grow the tree? – David Apr 9 '16 at 19:10
• I know the difference between Supervised and Unsupervised learning. As I said, I guess the Tree is grown according to some kind of Cost function in the Unsupervised case (Probably iterative method). I asked where did you see such thin so we could assist. Thank You. – Royi Apr 10 '16 at 5:55
• Tree-based clustering is probably the same as hierarchial clustering (en.wikipedia.org/wiki/Hierarchical_clustering) – Olli Niemitalo Apr 10 '16 at 19:29