Detecting Differentially Variable MicroRNAs via Model-Based Clustering

Figure 3

The boxplots of the 100 estimated Jaccard indices based on the 100 simulated datasets in SimIII (generating data from a mixture of multivariate normal distributions with zero marginal correlations). The closer the Jaccard index is to one, the better the performance of the method is. The closer the FPR (FNR) is to zero, the better the performance of the method is.

(a)

(b)

(c)