Supervised Clustering v.s. Unsupervised Clustering

The traditional clustering techniques tries to assign data points to different groups based on their similarity or common property measured in a high dimensional space. The testing criteria used to determine if a data points belongs to a certain cluster is generated collectively by the value of element in all dimensions for all data points. This approach resulted in an unsupervised clustering technique in which the criteria is determined by the data-set, this data-set based criteria may not always meets the needs of the problem on focus.