Degree-based random walk approach for graph embedding

Graph embedding, representing local and global neighbourhood information by numerical vectors, is a crucial part of the mathematical modeling wide range real-world systems. Among embedding algorithms, random walk-based algorithms have proven to be very successful. These collect creating numerous walks with predefined number steps. Creating most demanding process. The computation demand increases size network. Moreover, for networks, considering all nodes on same footing, abundance low-degree creates an imbalanced data problem. In this work, computationally less intensive node connectivity aware uniform sampling method proposed. proposed method, created proportionally degree node. advantages algorithm become more enhanced when applied large graphs. A comparative study using two namely CORA CiteSeer, presented. Compared fixed case, requires approximately 50% computational effort reach accuracy classification link prediction calculations.