Integrating social influence modeling and user modeling for trust prediction in signed networks

Trust and distrust between online users play an important role in social network applications, especially in the security domain. For example, trust information can enhance social recommendation and distrust information can be used for fraud detection. However, trust prediction is challenging due to the existence and imbalance of the three kinds of social status in signed social networks (i.e., trust, distrust and no-relation). Furthermore, there are a variety types of no-relation status in reality, e.g., strangers and frenemies, which cannot be well distinguished from the other social status by existing approaches. In this paper, we propose a novel Framework of Integrating both Latent and Explicit features (FILE), to better deal with the no-relation status and hence improve the overall trust/distrust prediction performance. In particular, we design two latent features to model user's intrinsic personality. Meanwhile, we design explicit features by extending social theories, to model the external social influence from mutual neighbors. The proposed model learns the features for each user via matrix factorization with a specially designed ranking-oriented loss function. Experimental results demonstrate the superior of our approach over the state-of-the-art methods, and the effectiveness of our approach in security applications. Our work sheds light on trust prediction in signed networks as well as security applications like fraud detection.