Following are the difference between K-Means and Hierarchical Clustering Algorithm (HCA)

- K-Means is that it needs us to pre-enter the number of clusters (K) but Hierarchical clustering has no such requirements. The algorithm on itself deduces the optimum number of cluster and displays it form of dendrogram.
- Performance of K-Means on spherical data is better than that of HCA
- Hierarchical clustering is a purely agglomerative approach and goes on to build one giant cluster. K-Means algorithm in all its iterations has same number of clusters.
- K-Means need circular data, while Hierarchical clustering has no such requirement.
- K-Means uses median or mean to compute centroid for representing cluster while HCA has various linkage method that may or may not employ the centroid.
- With introduction of mini batches K-Means can work with very large datasets but HCA lacks in this regard.
- Hierarchical methods are suited for cases which require arrangement of the clusters into a natural hierarchy. In K-means all clusters are on same level i.e. similar WCSS or cohesiveness.
- HCA can produce reproducible results while older versions of K-Means can’t
- K-Means simply divides data into mutually exclusive subsets while HCA arranges it into a tree format.