TriNE: Network Representation Learning for Tripartite Heterogeneous Networks
This work addresses representation learning for tripartite networks, common in real-world applications like user-item-tag systems, but it appears incremental as it extends existing embedding techniques to a specific network type.
The paper tackles the problem of learning node representations for tripartite heterogeneous networks, which involve three types of node entities, by developing TriNE, a method that models explicit and implicit relationships through metapath-guided random walks and a heterogeneous skip-gram model, resulting in validated performance for online user response prediction.
In this paper, we study network representation learning for tripartite heterogeneous networks which learns node representation features for networks with three types of node entities. We argue that tripartite networks are common in real world applications, and the essential challenge of the representation learning is the heterogeneous relations between various node types and links in the network. To tackle the challenge, we develop a tripartite heterogeneous network embedding called TriNE. The method considers unique user-item-tag tripartite relationships, to build an objective function to model explicit relationships between nodes (observed links), and also capture implicit relationships between tripartite nodes (unobserved links across tripartite node sets). The method organizes metapath guided random walks to create heterogeneous neighborhood for all node types in the network. This information is then utilized to train a heterogeneous skip-gram model based on a joint optimization. Experiments on real-world tripartite networks validate the performance of TriNE for the online user response prediction using embedding node features.