In example implementations, various interface functions are provided. In the Network View 300, an analyst can reorder the rows and columns of the matrix with certain criteria such as the node label, the average prediction score, and the total number of detected missing links. An analyst can also filter the matrix based on the prediction score, for example, to reveal the most probable missing links suggested by the algorithm. Moreover, different link prediction algorithms can be applied and viewed in the visualization to allow for an easy comparison of the results.

Further, an analyst can explore the link prediction results and add certain missing links to examine their influence with visual analysis of motifs and metrics described in the following. The added links are marked (e.g., as black crosses) on the matrix and also displayed at the top of the list. Individual links or a group of links can be added at once by selecting them from the matrix.