Skip to main content

Table 1 Comparison of biological coherence of clusters obtained for dataset I by different clustering algorithms

From: Microarray data mining: A novel optimization-based approach to uncover biologically coherent structures

    Proportion of Genes (%) in Clusters of p-values
   Average Correlation <= 10-4 <= 10-3
(Clustering Method) EP_GOS_Clust 0.617 32.8* 64.9*
  Iterated EP_GOS_Clust 0.685* > 69* > 90*
  KMedians 0.615 30.8 62.2*
  KCityBlk 0.398 27.5 56.7
  KCorr 0.630* 32.6* 60.1
  KMeans 0.614 25.1 55.2
  KAvePair 0.567 25.2 54.4
  QTClust 0.572 31.1 56.9
  SOTA 0.604 30.2 58.9
  SOM 0.623* 30.5 59.2
  1. Shown are the comparative biological coherence formed by various clustering methods on dataset I. The 2 shaded rows represent clustering done by the standalone EP_GOS_Clust backbone and the proposed iterative approach.
  2. * The top three performers in each category are indicated with an asterisk.