Figure 5
From: CLAG: an unsupervised non hierarchical clustering algorithm handling biological data
CLAG on brain cancer gene expression data: error analysis. Error analysis of CLAG clustering for gene expression data on brain cancer [18]. Data points are organized in five different pathologies and come from 42 patients. A: count of errors and key aggregates at increasing Δvalues, computed on all affine clusters (that is, with scores ≥ 0). Errors count both misclassified and unclustered patients. Notice that for all points plotted at Δ ≥ 0.1, the number of clustered patients is maximal, that is 42 (see B). B: number of clustered patients evaluated on aggregation of clusters having scores greater than a fixed threshold. C: number of PNET patients aggregated at increasing Δvalues, for different thresholds. Curves show that PNET patients aggregate slowly since they belong to clusters with low environmental and symmetric scores. D: number of patients that are correctly classified together.