Review on Learning and Extracting Graph Features for Link Prediction

Abstract

Link prediction in complex networks has attracted considerable attention from interdisciplinary research communities, due to its ubiquitous applications in biological networks, social networks, transportation networks, telecommunication networks, and, recently, knowledge graphs. Numerous studies utilized link prediction approaches in order sto find missing links or predict the likelihood of future links as well as employed for reconstruction networks, recommender systems, privacy control, etc. This work presents an extensive review of state-of-art methods and algorithms proposed on this subject and categorizes them into four main categories: similarity-based methods, probabilistic methods, relational models, and learning-based methods. Additionally, a collection of network data sets has been presented in this paper, which can be used in order to study link prediction. We conclude this study with a discussion of recent developments and future research directions.

An Integrative Network Science and Artificial Intelligence Drug Repurposing Approach for Muscle Atrophy in Spaceflight Microgravity

“…Once the DDKG is constructed any machine learning approach for link prediction can be used for predicting probable drugs. The network feature extraction method used here is based on random walks, which can be replaced with other local or global graph similarity based indices such as common neighbors, Jaccard index, Sorensen index, preferential attachment, Adamic-Adar index, resource allocation index, hub promoted index, Leicht-Holme-Newman index, parameter dependent index, local affinity structure index, individual attraction index, mutual information index, functional similarity weight, and local neighbors link index, Katz index, and page rank ( Fire et al, 2011 ; Mutlu et al, 2020 ). In this case, we found random walk and preferential attachment to give better results than the other features.…”

A Bayesian Graph Embedding Model for Link-Based Classification Problems

“…We then discuss the link prediction algorithms. The topology-based link prediction algorithms can be roughly classified into four categories: likelihood-based methods, probabilistic methods, graph embedding methods, and similaritybased methods [27]. Typical examples of the maximum likelihood-based methods are the stochastic block model (SBM) [28], fast probability block model (FBM) [29], hierarchical structure model (HSM) [30], [31].…”

“…Computing P (T ij = 1|θ(µ i , µ j )) is a binary classification problem, while estimating P(A ij = 1|E) is a typical link prediction problem. Furthermore, the interaction prediction problem is divided into a supervised machine learning task and an unsupervised machine learning task [27]. The difficulty lies in that there is no suitable way to integrate supervised and unsupervised learning models into the same framework.…”

Experimental Evaluation of Train and Test Split Strategies in Link Prediction

“…As discussed in the introduction, in this work we use the well-established similarity-based approach, where the feature for each pair of nodes x i = (u, v) consists of a particular score for each feature S feature (u, v). These scores are based solely on topological properties intrinsic to the network itself and not on any contextual information [12,14]. Hence, features used can be employed in any network, without requirements on node information available.…”