Structure-aware Hybrid-order Similarity Learning for Multi-view Unsupervised Feature Selection

2025-11-27

AI Summary

Existing multi-view unsupervised feature selection (MUFS) methods predominantly rely on first-order similarity graphs to model local structures, neglecting the global topological information encoded in second-order similarities; moreover, pre-defined second-order graphs are highly susceptible to noise. To address these limitations, this paper proposes a structure-aware hybrid-order similarity learning framework. It is the first to adaptively learn second-order similarity graphs within MUFS, while constructing cross-view relationships via consensus anchor points. This enables joint optimization and fusion of first-order (local proximity) and second-order (global topology) similarities. The resulting framework achieves both robustness against noise and strong discriminative capability. Extensive experiments on multiple real-world multi-view datasets demonstrate that our method significantly outperforms state-of-the-art approaches, achieving superior clustering accuracy and feature selection effectiveness.

Technology Category

• Machine Learning: Multi-instance/Multi-view Learning
• Computer Vision: Multi-modal Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Multi-view unsupervised feature selection (MUFS) has recently emerged as an effective dimensionality reduction method for unlabeled multi-view data. However, most existing methods mainly use first-order similarity graphs to preserve local structure, often overlooking the global structure that can be captured by second-order similarity. In addition, a few MUFS methods leverage predefined second-order similarity graphs, making them vulnerable to noise and outliers and resulting in suboptimal feature selection performance. In this paper, we propose a novel MUFS method, termed Structure-aware Hybrid-order sImilarity learNing for multi-viEw unsupervised Feature Selection (SHINE-FS), to address the aforementioned problem. SHINE-FS first learns consensus anchors and the corresponding anchor graph to capture the cross-view relationships between the anchors and the samples. Based on the acquired cross-view consensus information, it generates low-dimensional representations of the samples, which facilitate the reconstruction of multi-view data by identifying discriminative features. Subsequently, it employs the anchor-sample relationships to learn a second-order similarity graph. Furthermore, by jointly learning first-order and second-order similarity graphs, SHINE-FS constructs a hybrid-order similarity graph that captures both local and global structures, thereby revealing the intrinsic data structure to enhance feature selection. Comprehensive experimental results on real multi-view datasets show that SHINE-FS outperforms the state-of-the-art methods.

Problem

• Learns hybrid-order similarity graphs to capture local and global data structures
• Generates low-dimensional representations to identify discriminative features across views
• Enhances unsupervised feature selection by revealing intrinsic data structure

Innovation

• Learns consensus anchors and anchor graph for cross-view relationships
• Generates low-dimensional representations to reconstruct multi-view data
• Constructs hybrid-order similarity graph combining local and global structures