Dual-Information Driven Deep Multi-View Clustering for Heterogeneous Data

Deep multi-view clustering (DMVC) aims to utilize the consistency of multi-view data to learn a consensus representation using deep learning-based methods. However, existing methods overlook the presence of both semantic feature and topological structure information in the data. Also, the importance of these two information varies for heterogeneous data. To address these issues, we propose Dual-Information Driven Deep Multi-View Clustering for Heterogeneous Data (DID-DMVC). Firstly, to capture both semantic feature and topological structure information, we design a Dual-Information Extractor (DIE), which independently extracts two types of information. Secondly, we have designed a Tensor-Guided Low-Rank Fusion (TGLRF) strategy and developed a Dual-Level Adaptive Fusion Module (DLAFM). It adapts the fusion process by considering both the importance between views and the importance of the two different types of information for heterogeneous data. Thirdly, we design an Auxiliary Contrastive Loss (ACL) to regulate discrepancies between semantic feature and topological structure information during the DLAFM process. Finally, experiments demonstrate our model’s applicability across various types of datasets.