What are some of the pros and cons of hierarchical clustering compared to K-Means?

Pros:

  • Do not have to specify the number of clusters before running the algorithm
  • Results are reproducible and not subject to randomness introduced by choice of initial centroids as in K-Means

Cons:

  • Requires computation of pairwise linkage matrix, which can be computationally expensive
  • Results can differ based on the linkage criteria used
  • Can be sensitive to noise in data