Clustering gene expression data using a diffraction‐inspired framework

Table 5 Comparison of the clustering results for the Pomeroy dataset

Algorithm	AEC	AVI	ACC (%)
Diffractive clustering	63.1	258.7	67.8
k‐means	42.7	173.6	48.9
Fuzzy c‐means	39.9	154.6	43.3
Hierarchical clustering (single)	49.8	256.2	56.7
Hierarchical clustering (centroid)	50.7	292.6	57.8
Self‐organising map	41.2	13.7	44.4
Gaussian Mixture Model	42.2	79.2	37.7
DBSCAN	36.6	46.4	33.3

The diffractive clustering algorithm outperforms the others in terms of accuracy, with the density‐based and model‐based clustering algorithms performing relatively poorly due to the excessive mixing between the five different classes.

ISSN: 1475-925X