Stationary Diffusion State Neural Estimation for Multiview Clustering

Although many graph-based clustering methods attempt to model the stationary diffusion state in their objectives, performance limits using a predefined graph. We argue that estimation of can be achieved by gradient descent over neural networks. specifically design Stationary Diffusion State Neural Estimation (SDSNE) exploit multiview structural graph information for co-supervised learning. explore how network specially unsupervised learning and integrate multiple graphs into unified consensus shared self-attentional module. The view-shared module utilizes structure learn view-consistent global Meanwhile, instead auto-encoder most networks, SDSNE uses strategy with supervise as loss function guides achieving state. With help module, we obtain which nodes each connected component fully connect same weight. Experiments on several datasets demonstrate effectiveness terms six evaluation metrics.