Abstract
Network embedding, aiming to learn low-dimensional representations of nodes in networks, is very useful for many vector-based machine learning algorithms and has become a hot research topic in network analysis. Although many methods for network embedding have been proposed before, most of them are unsupervised, which ignores the role of prior information available in the network. In this paper, we propose a novel method for network embedding using semi-supervised kernel nonnegative matrix factorization (SSKNMF), which can incorporate prior information and thus to learn more useful features from the network through introducing kernel methodology. Besides, it can improve robustness against noises by using the objective function based on L <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> ,1 norm. Efficient iterative update rules are derived to resolve the network embedding model using the SSKNMF, and the convergence of these rules are strictly proved from the perspective of mathematics. The results from extensive experiments on several real-world networks show that our proposed algorithm is effective and has better performance than the existing representative methods.
Highlights
With the rapid growth of various network mechanisms, network analysis has attracted much attention from researchers in extensive fields, such as node classification [1], node clustering (a.k.a. community detection) [2], link prediction [3] and visualization, etc., in which the research tasks are significantly dependent on how the networks are represented.A network is represented normally by an adjacency matrix, which is usually very sparse and suffers from overwhelming dimensionality
We propose a method for network embedding using semi-supervised kernel nonnegative matrix factorization (SSKNMF), and the work is summarized as follows:
In this paper, we propose a network embedding method using semi-supervised kernel Nonnegative matrix factorization (NMF) (SSKNMF), which specially focuses on how to use prior information available in the network to further improve the performance of network embedding
Summary
A network is represented normally by an adjacency matrix, which is usually very sparse and suffers from overwhelming dimensionality. It cannot capture more complex and higher-order structural relationships hidden in the network, which as a result makes many tasks of network analysis costly in computation and ineffective in performance. To deal with these problems, network embedding has emerged and become a popular solution. After new representations of nodes have been learned, many tasks of network analysis can be effectively carried out by using conventional vector-based machine learning algorithms, such as K-means and support vector machine (SVM). The existing works have demonstrated that effective network embedding methods can help to improve the performance of different tasks on network analysis [4]–[6]
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have