RGCLN: Relational Graph Convolutional Ladder-Shaped Networks for Signed Network Clustering

Node embeddings are increasingly used in various analysis tasks of networks due to their excellent dimensional compression and feature representation capabilities. However, most researchers’ priorities have always been link prediction, which leads signed network clustering being under-explored. Therefore, we propose an asymmetric ladder-shaped architecture called RGCLN based on multi-relational graph convolution that can fuse deep node features generate representations with great representational power. adopts a framework capture convey information instead using the common method networks—balance theory. In addition, adds size constraint loss function prevent image-like overfitting during unsupervised learning process. Based learned by this end-to-end trained model, performs community detection large number real-world generative networks, results indicate our model has advantage over state-of-the-art embedding algorithms.