Abstract
The main task of network representation learning is to learn low-dimensional dense vector representations for nodes in the network, and to provide efficient data support for network analysis tasks such as node classification, link prediction, and node clustering. The existing network representation learning methods mainly include structure representation learning and attribute representation learning. However, researchers usually only consider the characteristics of regular nodes when modeling the structural and attribute characteristics of the network, ignoring the importance of edge nodes in the network. Although the number of edge nodes in the network is often lower than that of conventional nodes, they play a vital role in solving network analysis problems. In response to the above problems, this paper proposes an attribute network representation learning method ANE_BN for edge node discovery. This method applies structural representation learning and attribute representation learning methods to the representation learning problem of edge nodes for the first time, not only considering edge nodes attribute modeling also considers how to integrate attribute information and structural information into the node representation learning model. Specifically, first, the similarity of nodes in network structure is calculated by fusing high-level network structure information, and the similarity of nodes in attributes is calculated by cosine similarity, which provides data support for structure representation and attribute representation learning. Then, the non-negative matrix factorization method is used to model the representation learning model of the node in both the network structure and the attribute, and the relationship between the structural feature and the attribute feature is modeled by introducing the incidence matrix, so as to achieve structural stability and real-time attributes complementary advantages between sexes. In the experiment, the effectiveness of the algorithm ANE_BN proposed in this paper is verified through a real standard data set and compared with the existing classic algorithm. The experimental results show that the ANE_BN method can be better under the premise of ensuring the effectiveness of conventional node representation learning. The improved edge node indicates the effectiveness in various network analysis applications.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have