Fig. 5
From: On triangle inequalities of correlation-based distances for gene expression profiles
An example for average linkage hierarchical clustering with \(d_a\) and \(d_r\). A. The circle networks show the pairs where the distance is different in the ranks generated by \(d_a\) and \(d_r\) in step 2 to step 6. Nodes in network refer to samples for clustering. Edges refer to the distance where two samples are different in ranks in \(d_a\) and \(d_r\). \(c_1\) refers to the class generated in step 1, \(c_2\) refers to the class generated in step 2. B. Dendrogram for \(d_a\). C. Dendrogram for \(d_r\). The difference between the two dendrograms is colored in red. D. Distribution of ranks which are different in \(d_a\) and \(d_r\). E. Zoom in for the top 100 rank for D