Multi-view network embedding with node similarity ensemble

  • PDF / 725,601 Bytes
  • 16 Pages / 439.37 x 666.142 pts Page_size
  • 56 Downloads / 252 Views

DOWNLOAD

REPORT


Multi-view network embedding with node similarity ensemble Weiwei Yuan 1,2 & Kangya He 1 & Chenyang Shi 1 & Donghai Guan 1 & Yuan Tian 3 & Abdullah Al-Dhelaan 4 & Mohammed Al-Dhelaan 4 Received: 18 March 2019 / Revised: 2 November 2019 / Accepted: 6 February 2020 # Springer Science+Business Media, LLC, part of Springer Nature 2020

Abstract

Node similarity is utilized as the most popular guidance for network embedding: nodes more similar in a network should still be more similar when mapping node information from a high-dimensional vector space to a low-dimensional vector space. Most existing methods preserve a single node similarity in the network embedding, which can merely preserve one-side network structural information. Though some works try to utilize several node similarities to preserve more network information, they fail to consider the interrelationships between the latent spaces preserving different node similarities. This causes both network information insufficiency and network information redundancy. To solve the problems of existing works, we propose a novel multi-view network embedding model with node similarity ensemble. Node similarities are first selected to maximize the represented network information while minimizing the information redundancy. For each combination of the selected node similarities, a latent space is generated as a view of the network. A Canonical Correlation Analysis based approach is then used to extract the common structure of the latent spaces alignment, and a neural network based approach is used to extract the view-specific latent structure by measuring the asymmetric KL divergence of nodes’ Gaussian distribution. The common structure and the viewspecific structure of multiple views are merged to perverse the overall network information. Experiments held on the real-world networks verified the superiority of the proposed method to existing works. Keywords Node similarity . Multi-view learning . Network embedding

This article belongs to the Topical Collection: Special Issue on Application-Driven Knowledge Acquisition Guest Editors: Xue Li, Sen Wang, and Bohan Li

* Donghai Guan [email protected] Extended author information available on the last page of the article

World Wide Web

1 Introduction Learning vector representations for nodes in large networks has been proved extremely useful in various complex network analysis tasks [7, 9]. The main idea of network representation learning is to use dimensionality techniques to map the high-dimensional information about a node’s neighbors or other inherent properties into a low dimensional space [7, 9, 12, 22]. Since machine learning techniques cannot be implemented on the network directly, the learnednode representation is essential to machine learning based graph mining tasks such as node classification, link prediction or community detection [5, 7, 9]. Real world networks can be classified into three categories based on the structure: naive networks, attributes networks and multi-layer networks. The naive network only has