Struc2vec is a novel framework for learning latent representations of nodes based on their structural identity in networks. It uses a hierarchical approach to measure node similarity at different scales and constructs a multilayer graph to encode structural similarities and generate structural context for nodes. The framework generates random contexts for nodes using a weighted random walk on the multilayer graph, allowing for the learning of latent representations that capture structural identity. Numerical experiments show that struc2vec outperforms state-of-the-art techniques like DeepWalk and node2vec in capturing structural identity, especially in classification tasks where node labels depend more on structural identity than on homophily. Struc2vec is also robust to structural noise, such as random edge removal, and can be applied to large networks. The framework is flexible and can be used for various network analysis tasks, including identifying structural roles of nodes and classifying nodes based on their structural properties. The results demonstrate that struc2vec provides a more accurate and effective way to capture structural identity in networks compared to existing methods.Struc2vec is a novel framework for learning latent representations of nodes based on their structural identity in networks. It uses a hierarchical approach to measure node similarity at different scales and constructs a multilayer graph to encode structural similarities and generate structural context for nodes. The framework generates random contexts for nodes using a weighted random walk on the multilayer graph, allowing for the learning of latent representations that capture structural identity. Numerical experiments show that struc2vec outperforms state-of-the-art techniques like DeepWalk and node2vec in capturing structural identity, especially in classification tasks where node labels depend more on structural identity than on homophily. Struc2vec is also robust to structural noise, such as random edge removal, and can be applied to large networks. The framework is flexible and can be used for various network analysis tasks, including identifying structural roles of nodes and classifying nodes based on their structural properties. The results demonstrate that struc2vec provides a more accurate and effective way to capture structural identity in networks compared to existing methods.