Table 1. Cluster similarity to hand labels for two Bengalese finch and one Cassin’s vireo dataset.

Four clustering methods were used: (1) KMeans on spectrograms (2) KMeans on UMAP projections (3) HDBSCAN on first 100 principal components of spectrograms (4) HDBSCAN clustering of UMAP projections. With KMeans ‘K’ was set to the correct number of clusters to make it more competitive with HDBSCAN clustering. Standard deviation across individual birds is shown for the finch datasets. Best performing method for each metric is bolded.

	Homogeneity	Completeness	V-measure
B. Finch (Koumura)
KMeans	0.911±0.044	0.85±0.064	0.879±0.051
KMeans/UMAP	0.842±0.116	0.796±0.145	0.817±0.132
HDBSCAN/PCA	0.968±0.036	0.86±0.14	0.902±0.086
HDBSCAN/UMAP	0.99±0.006	0.74±0.122	0.841±0.088
B. Finch (Nicholson)
KMeans	0.954±0.024	0.707±0.101	0.809±0.074
KMeans/UMAP	0.967±0.018	0.688±0.098	0.801±0.072
HDBSCAN/PCA	0.901±0.067	0.837±0.027	0.866±0.034
HDBSCAN/UMAP	0.963±0.022	0.855±0.076	0.903±0.042
Cassin’s vireo
KMeans	0.894	0.808	0.849
KMeans/UMAP	0.928	0.829	0.875
HDBSCAN/PCA	0.849	0.906	0.877
HDBSCAN/UMAP	0.936	0.94	0.938