关键词: 符号网络, 局部社团, 结构平衡理论, 链路预测

Abstract: With the rapid development of the Internet industry, a bipartite signed network has become a commonly seen complex network. Based on the local community paradigm of the traditional unsigned networks and structural balance theory of signed networks, this paper proposes a local community paradigm for a bipartite signed network (BSLCP) for the first time. This theory considers the information of common neighbors in a signed network, as well as the links between them. The weighted balanced cycle increment ( WBCI) in a signed network is derived from BSLCP, which can represent the signed relationship between user nodes and product nodes in a bipartite signed network. On this basis, this paper further proposes the weighted balanced cycle increment classifier ( WBCIC) to predict links in the network. The test results show that WBCIC has better prediction ability than classical methods in link prediction.

Key words: Signed Network, Local Community, Structural Balance Theory, Link Prediction