The aim in clustering in particular, or unsupervised learning in general, is to find structure in the data. In the case of supervised learning (for example, in classification), this structure is imposed by the supervisor who defines the different classes and labels the instances in the training data by these classes.
