Diversity and consistency embedding learning for multi-view subspace clustering

With the emergence of multi-view data, many multi-view clustering methods have been developed due to the effectiveness of exploiting the complementary information of multi-view data. However, most existing multi-view clustering methods have the following two drawbacks: (1) they usually explore the r...

Full description

Saved in:
Bibliographic Details
Published in:Applied intelligence (Dordrecht, Netherlands) Netherlands), 2021-10, Vol.51 (10), p.6771-6784
Main Authors: Mi, Yong, Ren, Zhenwen, Mukherjee, Mithun, Huang, Yuqing, Sun, Quansen, Chen, Liwan
Format: Article
Language:English
Subjects:
Affinity
Artificial Intelligence
Clustering
Computer Science
Consistency
Embedding
Learning
Machines
Manufacturing
Mechanical Engineering
Methods
Optimization
Outliers (statistics)
Processes
Representations
Subspace methods
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:With the emergence of multi-view data, many multi-view clustering methods have been developed due to the effectiveness of exploiting the complementary information of multi-view data. However, most existing multi-view clustering methods have the following two drawbacks: (1) they usually explore the relationships between samples in the original space, where the high-dimensional features contain noise and outliers; (2) they only pay attention to exploring the consistency or enhancing the diversity of different views, such that the multi-view information cannot be completely utilized. In this paper, we propose a novel multi-view subspace clustering method, namely Diversity and Consistency Embedding Learning (DCEL), which learns a better affinity matrix in a learned latent embedding space while simultaneously considering diversity and consistency of multi-view data. Specifically, by leveraging a projection method, the multi-view data in the latent embedding space can be learned. Then, with the self-expression property, we seek a shared consistent representation among all views and a set of diverse representations of each view to better learn an affinity matrix in the latent embedding space. Furthermore, we develop an optimization scheme based on the alternating direction method of multipliers (ADMM) to solve the proposed method. Experimental evaluations on five benchmark datasets show the superiority of our method, compared with two single-view clustering methods and some state-of-the-art multi-view clustering methods.
ISSN:0924-669X
1573-7497
DOI:10.1007/s10489-020-02126-z