Multi-View Graph Clustering by Adaptive Manifold Learning

Graph-oriented methods have been widely adopted in multi-view clustering because of their efficiency learning heterogeneous relationships and complex structures hidden data. However, existing are typically investigated based on a Euclidean structure instead more suitable manifold topological structure. Hence, it is expected that will be to carry out intrinsic similarity learning. In this paper, we explore the implied adaptive for graph clustering. Specifically, our model seamlessly integrates multiple graphs into consensus with considered. We further manipulate useful rank constraint so its connected components precisely correspond distinct clusters. As result, able directly achieve discrete result without any post-processing. terms results, method achieves best performance 22 24 cases four evaluation metrics six datasets, which demonstrates effectiveness proposed model. computational performance, optimization algorithm generally faster or line other state-of-the-art algorithms, validates algorithm.