In this article, we will discuss hierarchical clustering algorithms in unsupervised machine learning. This algorithm is based on the splitting and merging of nested clusters. The linkage criteria to merge the clusters based on distance metric as shown below with a bottom-up approach. The linkage criteria give different clusters at different time speeds. Single linkage is not good in noisy data and ward linkage does not give proper cluster because the distance is not varied thus but good in properly balanced clusters and if we do not consider euclidean then average linkage can be used for clustering. The next parameter comes is connectivity that connects or merges the clusters based on the connectivity matrix.