Clustering is concerned with grouping instances into clusters. Instances in one cluster should be similar to each other and dissimilar to instances in other clusters. Clustering uses unlabeled data and, hence, requires an unsupervised learning technique.