Abstract

Many real-world datasets can be naturally described by multiple views. Due to this, multi-view learning has drawn much attention from both academia and industry. Compared to single-view learning, multi-view learning has demonstrated plenty of advantages. Clustering has long been serving as a critical technique in data mining and machine learning. Recently, multi-view clustering has achieved great success in various applications. To provide a comprehensive review of the typical multi-view clustering methods and their corresponding recent developments, this chapter summarizes five kinds of popular clustering methods and their multi-view learning versions, which include k-means, spectral clustering, matrix factorization, tensor decomposition, and deep learning. These clustering methods are the most widely employed algorithms for single-view data, and lots of efforts have been devoted to extending them for multi-view clustering. Besides, many other multi-view clustering methods can be unified into the frameworks of these five methods. To promote further research and development of multi-view clustering, some popular and open datasets are summarized in two categories. Furthermore, several open issues that deserve more exploration are pointed out in the end.

Highlights

  • Clustering is one of the most critical unsupervised learning techniques, which has been widely applied for data analysis, such as social network analysis, gene expression analysis, heterogeneous data analysis, and market analysis

  • Since covering all the proposed methods in one chapter is hard, to provide a comprehensive review of the typical multi-view clustering methods and their corresponding recent developments, we summarize five kinds of popular clustering methods and their multiview learning versions, which include k-means, spectral clustering, matrix factorization, tensor decomposition, and deep learning

  • We provide a comprehensive review of the typical multi-view clustering methods and their corresponding recent developments by focusing on five most typical and popular clustering methods, which include k-means, spectral clustering, matrix factorization, tensor decomposition, and deep learning

Read more

Summary

Introduction

Clustering is one of the most critical unsupervised learning techniques, which has been widely applied for data analysis, such as social network analysis, gene expression analysis, heterogeneous data analysis, and market analysis. To take better advantage of the multi-view information, the ideal approach is to simultaneously perform the clustering using each view of data features and integrate their results based on their importance to the clustering task. Since covering all the proposed methods in one chapter is hard, to provide a comprehensive review of the typical multi-view clustering methods and their corresponding recent developments, we summarize five kinds of popular clustering methods and their multiview learning versions, which include k-means, spectral clustering, matrix factorization, tensor decomposition, and deep learning This is based on the consideration that these clustering methods are the most widely employed algorithms for single-view data, and lots of efforts have been devoted to extending them for multi-view clustering.

Benefits of multi-view clustering
Benefit one: accurate description of data
Benefit two: reducing noises of data
Benefit three: wider range of applications
Multi-view clustering methods
Preliminaries of k-means
Basic form of multi-view k-means
Variants of multi-view k-means
Multi-view clustering via spectral clustering
Preliminaries of spectral clustering
Basic form of multi-view spectral clustering
Variants of multi-view spectral clustering
Multi-view clustering via matrix factorization
Preliminaries of matrix factorization
Basic form of multi-view matrix factorization
Variants of multi-view matrix factorization
Multi-view clustering via tensor decomposition
Preliminaries of tensor decomposition
Notations Let X be an m-order tensor of size I1 Â I2 Â ⋯ Â Im
CP decomposition
Tucker decomposition
Tensor decomposition-based multi-view clustering
Total variation based CP (TVCP)
Relations between Tucker decomposition and spectral clustering
Multi-view clustering via deep learning
Deep auto-encoder
Deep matrix factorization
Open datasets
Feature-based datasets
Graph-based datasets
Performance on different datasets
Open issues
View construction
Incomplete view
Single-view to multi-view
Deep leaning in multi-view
Conclusion
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call