Skip to main content
Figure 3 | EPJ Data Science

Figure 3

From: Prediction of new scientific collaborations through multiplex networks

Figure 3

ROC curves corresponding to dyadic metrics, AA single-layered scores, and the MAA metric. In solid lines, MC, CK and MAA scores. The MAA metric obtained with coefficients \(\eta _{ck}=0.05\) and \(\eta _{cr}=0.1\) shows the best performance of the whole set of metrics. The single-layer versions of the AA metric are shown in dashed (collaboration layer), dot-dashed (citation layer) and dotted (keywords layer) lines. The normalized metrics (NMC and NCK) completely overlap their non-normalized counterparts, and thus are removed for clarity. Similarly, the AA metric applied over the aggregated network (not shown) completely overlaps with the \(\mathit{AA}_{k}\) curve, showing that the main contribution in the aggregated network comes from the keywords layer

Back to article page