Table 7 Accuracies (ACC) of the unsupervised clustering of the Temperature Dataset, for several parametric clustering algorithms, and several values of the pre-specified number of clusters. For each value of the number of clusters parameter, the unsupervised clustering accuracies are computed using the taxonomic labels as ground truth (top row), respectively the environment category labels as ground truth (bottom row).
From: Environment and taxonomy shape the genomic signature of prokaryotic extremophiles
No. Clusters | Labelling | Unsupervised clustering accuracy - ACC (%) | |||
|---|---|---|---|---|---|
K-means | K-medoids | GMM | iDeLUCS | ||
2 | Taxonomy | 63.84 | 63.92 | 63.23 | 68.97 |
Environment | 36.27 | 36.50 | 36.26 | 38.23 | |
4 | Taxonomy | 63.99 | 77.65 | 68.45 | 75.44 |
Environment | 34.37 | 40.81 | 38.30 | 48.31 | |
8 | Taxonomy | 87.81 | 82.13 | 77.79 | 81.48 |
Environment | 50.99 | 49.63 | 53.74 | 56.77 | |
