Learning the Consistent Behavior of Common Users for Target Node Prediction across Social Networks
Shan-Hung Wu, Hao-Heng Chien, Kuan‐Hua Lin, Philip S. Yu · 2014
We study the target node prediction prob-lem: given two social networks, identify those nodes/users from one network (called the source network) who are likely to join another (called the target network, with nodes called target nodes). Although this problem can be solved us-ing existing techniques in the field of cross do-main classification, we observe that in many real-world situations the cross-domain classifiers per-form sub-optimally due to the heterogeneity be-tween source and target networks that prevents the knowledge from being transferred. In this paper, we propose learning the consistent be-havior of common users to help the knowledge transfer. We first present the Consistent Inci-dence Co-Factorization (CICF) for identifying the consistent users, i.e., common users that be-have consistently across networks. Then we in-troduce the Domain-UnBiased (DUB) classifiers that transfer knowledge only through those con-sistent users. Extensive experiments are con-ducted and the results show that our proposal copes with heterogeneity and improves predic-tion accuracy. 1.