Graph Embedding Method Based on Biased Walking for Link Prediction

Link prediction is an essential and challenging problem in research on complex networks, which can provide tools theoretical supports for the formation evolutionary mechanisms of networks. Existing graph representation learning methods based random walks usually ignore influence local network topology transition probability walking nodes when predicting existence links, sampling strategy during uncontrolled, leads to inability these effectively learn high-quality node vectors solve link problem. To address above challenges, we propose a novel embedding method prediction. Specifically, analyze evolution mechanism links triadic closure theory use clustering coefficient represent aggregation ability network’s structure, this adaptive definition structure enables control process. Finally, generated biased paths employed Extensive experiments analyses show that TCW algorithm provides high accuracy across diverse set datasets.