Adaptive Similarity Function with Structural Features of Network Embedding for Missing Link Prediction

Link prediction is a fundamental problem of data science, which usually calls for unfolding the mechanisms that govern micro-dynamics networks. In this regard, using features obtained from network embedding predicting links has drawn widespread attention. Although methods based on edge or node similarity have been proposed to solve link problem, many technical challenges still exist due unique structural properties networks, especially when networks are sparse. From graph mining perspective, we first give empirical evidence inconsistency between heuristic and learned features. Then, propose novel framework, AdaSim, by introducing an Adaptive Similarity function random walks. The feature representations optimizing graph-based objective function. Instead generating binary operators, perform solely leveraging network. We define flexible with one tunable parameter, serves as penalty original measure. optimal value through supervised learning thus adaptive distribution. To evaluate performance our algorithm, conduct extensive experiments eleven disparate real world. Experimental results show AdaSim achieves better than state-of-the-art algorithms robust different sparsities