Link Prediction Techniques Research Articles

A Comprehensive Survey of Link Prediction Techniques for Social Network

A growing trend of using social networking sites is attracting researchers to study and analyze different aspects of social network. Besides many problems, link prediction is a fascinating problem in the field of social network analysis (SNA). Link prediction, in social network analysis, is a task

Link prediction in networks using effective transitions

We introduce a new method for predicting the formation or the detection of unobserved links in real-world networks, which we refer to as the method of effective transitions. This method relies on the theory of isospectral matrix reductions to compute the probability of eventually transitioning from one vertex to another in a (biased) random walk on the network. This link prediction technique is quite versatile and can be used to predict links in networks that are directed or undirected and are either weighted or unweighted. We apply this method to a large number of social, technological, and natural networks and show that it is competitive with other link predictors, often outperforming them. We also provide a method of approximating our effective transition method and show that aside from having much lower temporal complexity, this approximation often provides more accurate predictions than the original effective transition method. We also prove a number of results regarding our effective transition algorithm and its approximation.

A Novel Method to Dynamically Fix Threshold for Node Neighbourhood Based Link Prediction Techniques

The objective of an online social network is to amplify the stream of information among the users. This goal can be accomplished by maximizing interconnectivity among users using link prediction techniques. Existing link prediction techniques uses varied heuristics such as similarity score to predict possible connections. Link prediction can be considered a binary classification problem where probable class outcomes are presence and absence of connections. One of the challenges in classification is to decide threshold value. Since the social network is exceptionally dynamic in nature and each user possess different features, it is difficult to choose a static, common threshold which decides whether two non-connected users will form interconnectivity. This article proposes a novel technique, FIXT, that dynamically decides the threshold value for predicting the possibility of new link formation. The article evaluates the performance of FIXT with six baseline techniques. The comparative results depict that FIXT achieves accuracy up to 93% and outperforms baseline techniques.

Hybrid Feature Based Approach for Recommending Friends in Social Networking Systems

Link prediction is an effective technique to be applied on graph-based models due to its wide range of applications. It helps to understand associations between nodes in social communities. The social networking systems use link prediction techniques to recommend new friends to their users. In this paper, we design two time efficient algorithms for finding all paths of length-2 and length-3 between every pair of vertices in a network which are further used in computation of final similarity scores in the proposed method. Further, we define a hybrid feature-based node similarity measure for link prediction that captures both local and global graph features. The designed similarity measure provides friend recommendations by traversing only paths of limited length, which causes more faster and accurate friend recommendations. Experimental results show adequate level of accuracy in friend recommendations within considerable computing time.

A Similarity-Inclusive Link Prediction Based Recommender System Approach

Despite being a challenging research field with many unresolved problems, recommender systems are getting more popular in recent years. These systems rely on the personal preferences of users on items given in the form of ratings and return the preferable items based on choices of like-minded users. In this study, a graph-based recommender system using link prediction techniques incorporating similarity metrics is proposed. A graph-based recommender system that has ratings of users on items can be represented as a bipartite graph, where vertices correspond to users and items and edges to ratings. Recommendation generation in a bipartite graph is a link prediction problem. In current literature, modified link prediction approaches are used to distinguish between fundamental relational dualities of like vs. dislike and similar vs. dissimilar. However, the similarity relationship between users/items is mostly disregarded in the complex domain. The proposed model utilizes user-user and item-item cosine similarity value with the relational dualities in order to improve coverage and hits rate of the system by carefully incorporating similarities. On the standard MovieLens Hetrec and MovieLens datasets, the proposed similarity-inclusive link prediction method performed empirically well compared to other methods operating in the complex domain. The experimental results show that the proposed recommender system can be a plausible alternative to overcome the deficiencies in recommender systems.

Analyzing the Bills-Voting Dynamics and Predicting Corruption-Convictions Among Brazilian Congressmen Through Temporal Networks

In this paper, we propose a network-based technique to analyze bills-voting data comprising the votes of Brazilian congressmen for a period of 28 years. The voting sessions are initially mapped into static networks, where each node represents a congressman and each edge stands for the similarity of votes between a pair of congressmen. Afterwards, the constructed static networks are converted to temporal networks. Our analyses on the temporal networks capture some of the main political changes happened in Brazil during the period of time under consideration. Moreover, we find out that the bills-voting networks can be used to identify convicted politicians, who commit corruption or other financial crimes. Therefore, we propose two conviction prediction methods, one is based on the highest weighted convicted neighbor and the other is based on link prediction techniques. It is a surprise to us that the high accuracy (up to 90% by the link prediction method) on predicting convictions is achieved only through bills-voting data, without taking into account any financial information beforehand. Such a feature makes possible to monitor congressmen just by considering their legal public activities. In this way, our work contributes to the large scale public data study using complex networks.

Relative Assortativity Index: A Quantitative Metric to Assess the Impact of Link Prediction Techniques on Assortativity of Complex Networks

Abstract We propose a quantitative metric (called relative assortativity index, RAI) to assess the extent with which a real-world network would become relatively more assortative due to link addition(s) using a link prediction technique. Our methodology is as follows: for a link prediction technique applied on a particular real-world network, we keep track of the assortativity index values incurred during the sequence of link additions until there is negligible change in the assortativity index values for successive link additions. We count the number of network instances for which the assortativity index after a link addition is greater or lower than the assortativity index prior to the link addition and refer to these counts as relative assortativity count and relative dissortativity count, respectively. RAI is computed as (relative assortativity count − relative dissortativity count) / (relative assortativity count + relative dissortativity count). We analyzed a suite of 80 real-world networks across different domains using 3 representative neighborhood-based link prediction techniques (Preferential attachment, Adamic Adar and Jaccard coefficients [JACs]). We observe the RAI values for the JAC technique to be positive and larger for several real-world networks, while most of the biological networks exhibited positive RAI values for all the three techniques.

Competition or collaboration? – Analysis of technological knowledge ecosystem within the field of alternative powertrain systems: A patent-based approach

In light of the transition towards a low-carbon economy and growing social demands for sustainable energy consumption, alternative powertrain systems, such as battery electric vehicles (BEVs), hybrid electric vehicles (HEVs) and fuel cell electric vehicles (FCEVs), have gained significance as a viable solution to revolutionize the dominant design logic of internal combustion engines (ICE) in automotive applications. By assuming that technological innovations emerge from the ability to recombine existing technologies, tracking and forecasting the technological knowledge interaction trajectory could assist experts and decision-makers in guiding their future R&D activities. To this end, a patent-based analysis framework is introduced to study the technological knowledge ecosystem of BEVs, HEVs and FCEVs in terms of technological knowledge flow (TKF). In this study, the principles of social network analysis are firstly used to quantify and visualize the connections among interacting technological knowledge areas. Secondly, link prediction technique is applied to predict the newly emerging and decaying links from a forward-looking perspective as well as the changing interrelationships among different technological knowledge areas. The findings show that the alternative powertrain systems form a cohesive technological knowledge ecosystem, pushing the underlying technological knowledge ecosystem to adopt a coopetition-based growth strategy. This study can provide valuable information for stakeholders interested in evaluating the evolving characteristics of alternative powertrain system technologies and offer an enhanced understanding on the technological competition in the automotive industry with regard to developing cleaner and more efficient mobility solutions.

A Recursive Bayesian Approach for the Link Prediction Problem

Recently, link prediction techniques have been increasingly adopted to discover link patterns in various domains. On challenging problem is to improve the performance continually. In this paper, we propose a recursive prediction mechanism to addresses the link prediction problem. A posterior is calculated based on observed data, and then we estimate the state of the graph and use the posterior as the prior distribution for the next stage. With the increasing of iterations, the proposed approach incorporates more and more topological structure information and node attributes data. Experimental results with real-world networks have shown that the proposed solution performs better in terms of well-known metrics as compared to the existing approaches. This novel approach has already been integrated into an expert system and provides auxiliary support for decision-makers.

A new classification algorithm recommendation method based on link prediction

Recommending appropriate classification algorithm(s) for a given classification problem is of great significance and also one of the challenging problems in the field of data mining, which is usually viewed as a meta-learning problem. Multi-label learning has been adopted and validated to be an effective meta-learning method in classification algorithm recommendation. However, the multi-label learning method used in previous classification algorithm recommendation relies only on relationship between data sets and their direct neighbours, ignoring the impact of other data sets. In this paper, a new classification algorithm recommendation method based on link prediction between data sets and classification algorithms is proposed. Taking advantage of link prediction in heterogeneous networks, this method considers the impact of all data sets and makes full use of the interactions between data sets as well as between data sets and algorithms. Firstly, meta data of the training data sets is collected. And then a heterogeneous network called DAR (Data and Algorithm Relationship) Network is constructed with the meta data. Finally, the link prediction technique is adopted to recommend appropriate algorithm(s) for a given data set on the basis of the DAR Network. To evaluate the proposed link prediction-based recommendation method, extensive experiments with 131 data sets and 21 classification algorithms are conducted. Results of 5 performance measures indicate that the proposed method is more effective compared with the base line classification algorithm recommendation method and can be used in practice.

Predicting potential drug-drug interactions on topological and semantic similarity features using statistical learning.

Drug-drug interaction (DDI) is a change in the effect of a drug when patient takes another drug. Characterizing DDIs is extremely important to avoid potential adverse drug reactions. We represent DDIs as a complex network in which nodes refer to drugs and links refer to their potential interactions. Recently, the problem of link prediction has attracted much consideration in scientific community. We represent the process of link prediction as a binary classification task on networks of potential DDIs. We use link prediction techniques for predicting unknown interactions between drugs in five arbitrary chosen large-scale DDI databases, namely DrugBank, KEGG, NDF-RT, SemMedDB, and Twosides. We estimated the performance of link prediction using a series of experiments on DDI networks. We performed link prediction using unsupervised and supervised approach including classification tree, k-nearest neighbors, support vector machine, random forest, and gradient boosting machine classifiers based on topological and semantic similarity features. Supervised approach clearly outperforms unsupervised approach. The Twosides network gained the best prediction performance regarding the area under the precision-recall curve (0.93 for both random forests and gradient boosting machine). The applied methodology can be used as a tool to help researchers to identify potential DDIs. The supervised link prediction approach proved to be promising for potential DDIs prediction and may facilitate the identification of potential DDIs in clinical research.

A new similarity measure for link prediction based on local structures in social networks

Link prediction is a fundamental problem in social network analysis. There exist a variety of techniques for link prediction which applies the similarity measures to estimate proximity of vertices in the network. Complex networks like social networks contain structural units named network motifs. In this study, a newly developed similarity measure is proposed where these structural units are applied as the source of similarity estimation. This similarity measure is tested through a supervised learning experiment framework, where other similarity measures are compared with this similarity measure. The classification model trained with this similarity measure outperforms others of its kind.

An Efficient Link Prediction Technique in Social Networks based on Node Neighborhoods

The unparalleled accomplishment of social networking sites, such as Facebook, LinkedIn and Twitter has modernized and transformed the way people communicate to each other. Nowadays, a huge amount of information is being shared by online users through these social networking sites. Various online friendship sites such as Facebook and Orkut, allow online friends to share their thoughts or opinions, comment on others’ timeline or photos, and most importantly, meet new online friends who were known to them before. However, the question remains as to how to quickly propagate one’s online network by including more and more new friends. For this, one of the easy methods used is list of ‘Suggested Friends’ provided by these online social networking sites. For suggestion of friends, prediction of links for each online user is needed to be made based on studying the structural properties of the network. Link prediction is one of the key research directions in social network analysis which has attracted much attention in recent years. This paper discusses about a novel efficient link prediction technique LinkGyp and many other commonly used existing prediction techniques for suggestion of friends to online users of a social network and also carries out experimental evaluations to make a comparative analysis among each technique. Our results on three real social network datasets show that the novel LinkGyp link prediction technique yields more accurate results than several existing link prediction techniques.

Prediction of Missing Links in a Social Networks: FINN (Feature Integration with Node Neighbor)

Link prediction techniques are used to identify the future network structure on the basis of existing connectivity pattern of the users. Most of the existing link prediction techniques employ varied similarity indices to predict new links in network. Some techniques use common neighbours while others use common shared profile information of the user for prediction. Typically existing link prediction techniques have only focused on one of these two data modalities: common neighbours or common attributes. Both of them play equally important role in the dynamics of the network. In this paper, we propose a feature integrated node neighbour (FINN) approach, an accurate algorithm for predicting links in network. FINN integrates Jaccard coefficient and Adamic Adar to predict link between nodes using their connections and features. We have evaluated FINN by implementing it over the real-time Facebook dataset collected from SNAP repository and validated the result through area under ROC curve.

Link Prediction Schemes Contra Weisfeiler-Leman Models

Link prediction is of particular interest to the data mining and machine learning communities. Until recently all approaches to the problem used embedding-based methods which leverage either node similarities or latent group memberships towards link prediction. Chen and Zhang recently developed a class of non-embedding approaches called Weisfeiler-Leman (WL) Models. WL-Models extract subgraphs around links and then encode subgraph patterns via adjacency matrices using the so-called Palette-WL algorithm. A training stage then learns nonlinear graph topological features for link prediction. Chen and Zhang compared two WL-Models – a linear regression model (“WLLR”) and a neural networks model (“WLNM”) – against 12 different common link prediction schemes. In this paper, all author claims are validated for WLLR. Additionally, WLLR is tested against 22 additional embedding-based link prediction techniques arising from common neighbor-, path- and random walk-based schemes. WLLR is shown not to be superior when calculable. In fact, in 80% of the datasets where comparisons were possible, one of our added implementations proved superior.

Link Prediction on Directed Networks Based on AUC Optimization

Complex networks have become high-dimensional, sparse, and redundant due to the rapid expansion of the Internet. Effective link prediction techniques are needed to obtain the most relevant and important information for Internet users. A new link prediction algorithm based on the area under the receiver operating characteristic curve (AUC) as the evaluation metric is proposed in this paper. In the proposed method, the AUC is treated as the objective function and the link prediction problem is transformed into an optimization problem. A group of topological features is defined for each ordered pair of nodes. By using those features as the attributes of the node pairs, link prediction can be treated as a binary classification, where the class label of each node pair is determined by the existence of a directed link between the node pair. Then, the binary classification problem can be solved by the AUC optimization. According to the empirical results, high-quality predictions can be achieved by our algorithm.

Link prediction based on non-negative matrix factorization.

With the rapid expansion of internet, the complex networks has become high-dimensional, sparse and redundant. Besides, the problem of link prediction in such networks has also obatined increasingly attention from different types of domains like information science, anthropology, sociology and computer sciences. It makes requirements for effective link prediction techniques to extract the most essential and relevant information for online users in internet. Therefore, this paper attempts to put forward a link prediction algorithm based on non-negative matrix factorization. In the algorithm, we reconstruct the correlation between different types of matrix through the projection of high-dimensional vector space to a low-dimensional one, and then use the similarity between the column vectors of the weight matrix as the scoring matrix. The experiment results demonstrate that the algorithm not only reduces data storage space but also effectively makes the improvements of the prediction performance during the process of sustaining a low time complexity.

Link-prediction to tackle the boundary specification problem in social network surveys.

Diffusion processes in social networks often cause the emergence of global phenomena from individual behavior within a society. The study of those global phenomena and the simulation of those diffusion processes frequently require a good model of the global network. However, survey data and data from online sources are often restricted to single social groups or features, such as age groups, single schools, companies, or interest groups. Hence, a modeling approach is required that extrapolates the locally restricted data to a global network model. We tackle this Missing Data Problem using Link-Prediction techniques from social network research, network generation techniques from the area of Social Simulation, as well as a combination of both. We found that techniques employing less information may be more adequate to solve this problem, especially when data granularity is an issue. We validated the network models created with our techniques on a number of real-world networks, investigating degree distributions as well as the likelihood of links given the geographical distance between two nodes.

Prediction of Datasets sameAs Interlinking on Web of Data

In order to be considered as Linked Data, the datasets on the web must be linked to other datasets. We focus on predicting the possible links between datasets with the most important RDF link type, owl:sameAs using link prediction and classification techniques. Since the goal is to discriminate between linked dataset pairs against not-linked ones, we formulate the link prediction problem as a classification problem. We adopt Random Forest as the basic classifier to incorporate features of the scores output by unsupervised predictors, and apply the bagging technique to combine multiple forests to reduce variance and improve the accuracy. Experiments show we can improve the prediction performance by about 10% in AUROC compared with the best unsupervised predictor.

Retweets as a Predictor of Relationships among Users on Social Media.

Link prediction is the problem of detecting missing links or predicting future link formation in a network. Application of link prediction to social media, such as Twitter and Facebook, is useful both for developing novel services and for sociological analyses. While most existing research on link prediction uses only the social network topology for the prediction, in social media, records of user activities such as posting, replying, and reposting are available. These records are expected to reflect user interest, and so incorporating them should improve link prediction. However, research into link prediction using the records of user activities is still in its infancy, and the effectiveness of such records for link prediction has not been fully explored. In this study, we focus in particular on records of reposting as a promising source that could be useful for link prediction, and investigate their effectiveness for link prediction on the popular social media platform Twitter. Our results show that (1) the prediction accuracy of techniques using reposting records is higher than that of popular topology-based techniques such as common neighbors and resource allocation for actively retweeting users, (2) the accuracy of link prediction techniques that use network topology alone can be improved by incorporating reposting records.