Figure 4
From: Integrating multi-platform genomic data using hierarchical Bayesian relevance vector machines
Boxplots of weights. Boxplots of weights β=(β1,β2,β3) of gene expressions, miRNA expressions, and their interactions in predicting log survival time. The GBM survival data is randomly split 50 times into training and test data, H-RVM is fit on the training data, and β is obtained from the fit on training data. The y-axis shows the distributions of respective weights for gene expressions, miRNA expressions, and their interactions in predicting log survival time of patients across 50 random splits of the GBM survival data.