Skip to main content

Table 4. The execution time before and after applying UMAP on the different clustering algorithms on the five datasets.

From: Considerably Improving Clustering Algorithms Using UMAP Dimensionality Reduction Technique: A Comparative Study

Time in second MNIST F-MNIST UMIST face Pen digits USPS
K-means 112.13 74.69 17.24 0.94 12.93
UMAP + K-means 1.22 1.20 0.33 0.26 0.57
Agglomerative 710.08 674.14 6.57 0.48 47.93
UMAP + Agglomerative 88.31 100.14 0.03 0.28 8.51
HDBSCAN 1603.26 1660.25 17.77 1.14 117.56
UMAP + HDBSCAN 5.13 4.49 0.03 0.12 0.75
GMM 24.51 26.27 3.49 0.58 25.26
UMAP+GMM 0.51 0.42 0.06 0.03 0.14