Skip to main content

Table 2. Comparison between the different clustering algorithms on the five datasets according to the accuracy measure.

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

  MNIST F-MNIST UMIST face Pen digits USPS
K-means 0.5278 0.4750 0.4348 0.7028 0.6678
UMAP + K-means 0.9054 0.5865 0.7409 0.8843 0.8105
Improvement score 0.3776 0.1115 0.3061 0.1815 0.1427
Agglomerative 0.5751 0.5766 0.4539 0.7451 0.6834
UMAP + Agglomerative 0.8918 0.5925 0.7270 0.8737 0.9584
Improvement score 0.3167 0.0159 0.2731 0.1286 0.2740
HDBSCAN 0.2765 0.2140 0.4904 0.5453 0.3529
UMAP + HDBSCAN 0.7765 0.3458 0.6730 0.9004 0.9553
Improvement score 0.5000 0.1318 0.1826 0.3551 0.6024
GMM 0.4507 0.4579 0.3826 0.4836 0.4802
UMAP+GMM 0.9159 0.5885 0.7287 0.8748 0.6727
Improvement score 0.4652 0.1306 0.3461 0.3912 0.1925