NETWORK EMBEDDING OF TOPIC-ATTENTION NETWORK BASED ON SET PAIR ANALYSIS

摘要

Concerning the problem that heterogeneous network embedding only considers social relations in structure and ignores semantics, combining the social relationship between users and the preference of users for topics, a network embedding algorithm based on topic-attention network was proposed. Firstly, according to the characteristics of the topic-attention network and combining with the idea of the identical-discrepancy-contrary (determination and uncertainty) of set pair analysis (SPA) theory, the transition probability model was given. Then a random walk algorithm based on two types of nodes was proposed by using the transition probability model, which obtained relatively high-quality random walk sequences. Finally, the embedding vector space representation of the topic-attention network was obtained by modeling based on two types of nodes in the sequences. After theoretical analysis, experimental results on the Douban dataset show that the modularity of the proposed algorithm is 0.5871 when the number of the overlapping communities is 11, which is nearly 6.5% higher than that of rnetapath2vec algorithm. The random walk algorithm combined with the transition probability model is more comprehensive in analyzing the connection relationship between nodes in the network, and can capture more detailed information in the network.