Figure 8

CLAG on the 128-dimensional synthetic dataset. A: the 128-dimensional dataset contains 1024 points and 16 clusters generated with a gaussian distribution (http://cs.joensuu.fi/sipu/datasets/DIM128.txt). CLAG perfectly distinguished the 16 clusters when run with Δ = 0.05and scores ≥ 0.5. B: curves associated to different score thresholds describing the number of elements that are clustered by CLAG while varying Δvalues. Note that the number of elements is 1011 for Δ = 0.05and maximal scores. Different clustering algorithms were run on this dataset: k-means (C), c-means (D), MCLUST (E). k-means and c-means were run with 16 clusters, and MCLUST with “ellipsoidal, equal variance with 9 components” as best model (note the 8 grey clusters). For k-means, clusters 1, 13 are split in several k-means clusters while clusters 3, 8 (violet) and 4, 16 (light blue) are fused together. c-means clusters the original ensemble in only 11 clusters: clusters 10, 4, 16 (brown) and 5, 6, 8, 9 (orange) are grouped together. In A, C, D, E elements are represented by circles. Different clusters are distinguished by different colors. Figures ACDE are realized by plotting the first two columns of the matrix describing the dataset.