Fast and accurate link prediction in social networking systems

Papadimitriou, Alexis; Symeonidis, Panagiotis; Manolopoulos, Yannis

doi:10.1016/j.jss.2012.04.019

Cited by 169 publications

(106 citation statements)

References 16 publications

Supporting

Mentioning

105

Contrasting

Unclassified

Order By: Relevance

“…Bounded Local Path Index (BLP) [29] bounds local paths with structural coefficients according to path lengths:…”

Section: Baselinesmentioning

confidence: 99%

“…In order to find a nice tradeoff between performance and complexity, the third class of similarity metrics were proposed on quasi-local structures. The Local Path Index (LP) ignored long-path terms in Katz Index [27,28], and its bounded version (BLP) relates local paths in an elaborate way [29]. The Local Random Walk Index (LRW) limits a random walker within a local range [30], while the Superposed Random Walk Index (SRW) continuously released a random walker at the starting node to emphasize the nodes near the target node [30].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Predicting missing links via significant paths

Zhu¹,

Tian²,

Cai³

et al. 2014

EPL

View full text Add to dashboard Cite

Link prediction plays an important role in understanding the intrinsic evolving mechanisms of networks. With the belief that the likelihood of the existence of a link between two nodes is strongly related to their similarity, many methods have been proposed to calculate node similarity based on node attributes and/or topological structures. Among a large variety of methods that take into account paths connecting the target pair of nodes, most of them neglect the heterogeneity of those paths. Our hypothesis is that a path consisting of small-degree nodes provides a strong evidence of similarity between two ends, accordingly, we propose a so-called significant path index in this Letter to leverage intermediate nodes' degrees in similarity calculation. Empirical experiments on twelve disparate real networks demonstrate that the proposed index outperforms the mainstream link prediction baselines.

show abstract

“…Bounded Local Path Index (BLP) [29] bounds local paths with structural coefficients according to path lengths:…”

Section: Baselinesmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Predicting missing links via significant paths

Zhu¹,

Tian²,

Cai³

et al. 2014

EPL

View full text Add to dashboard Cite

show abstract

“…We compare experimentally total seven link prediction techniques. In particular, we compare the Katz [12], the Adamic/Adar [20], the PA [9], the FOAF [10], the LA [11], the FriendLink [19], and the FriendTNS [18]. All the algorithms are implemented in R programming language.…”

Section: Data and Experimental Setupmentioning

confidence: 99%

A Survey on Link Prediction Problem in Social Networks

Kumar¹

2017

IJRASET

View full text Add to dashboard Cite

Link prediction is a crucial task in analysis of online social networks. It can be used for friend recommendations in social networks. It also has applications in other areas like bioinformatics, information retrieval and e-commerce. Currently, variety of techniques is available to deal with link prediction problem, ranging from local feature based approaches to global feature based approaches. Some recent techniques combine the advantages of these two approaches and comes under hybrid category. all these methods differ from each other with respect to prediction performance in terms of accuracy, efficiency, and generalization ability. In this paper, we survey some representative link prediction techniques under each category. We largely consider three types of approaches. First category is of local feature based approaches which do not exploit the whole network structure. Second category is of global feature based approaches which use the overall path structure of a network. And, finally the hybrid approaches which combine the advantages of above two approaches (local and global) in computation of similarity scores between each pair of nodes in a network. We discuss some recent existing approaches corresponding to these broad categories and analyze their strength and weaknesses. We conclude the paper with discussion on recent developments and future research direction.

show abstract

“…a network consisting of two separated node sets and links between the two node sets. Examples of this kind of networks are scientific collaboration network [3,4], P2P Internet formed by computer terminals and data [5]; cooperative network formed by actors and their films [6,7]; Shares Network formed by the investors and companies [8,9]; the activity network formed by the members of the club and their activities [10]; the audience with the songs network [11]; disease-gene networks [12], and so on.…”

Section: Introductionmentioning

confidence: 99%

Information filtering in evolving online networks

Chen

Zhang

et al. 2018

Physics Letters A

View full text Add to dashboard Cite

Recommender systems use the records of users' activities and profiles of both users and products to predict users' preferences in the future. Considerable works towards recommendation algorithms have been published to solve the problems such as accuracy, diversity, congestion, cold-start, novelty, coverage and so on. However, most of these research did not consider the temporal effects of the information included in the users' historical data. For example, the segmentation of the training set and test set was completely random, which was entirely different from the real scenario in recommender systems. More seriously, all the objects are treated as the same, regardless of the new, the popular or obsoleted products, so do the users. These data processing methods always lose useful information and mislead the understanding of the system's state. In this paper, we detailed analyzed the difference of the network structure between the traditional random division method and the temporal division method on two benchmark data sets, Netflix and MovieLens. Then three classical recommendation algorithms, Global Ranking method, Collaborative Filtering and Mass Diffusion method, were employed. The results show that all these algorithms became worse in all four key indicators, ranking score, precision, popularity and diversity, in the temporal scenario. Finally, we design a new recommendation algorithm based on both users' and objects' first appearance time in the system. Experimental results showed that the new algorithm can greatly improve the accuracy and other metrics.

show abstract

Fast and accurate link prediction in social networking systems

Cited by 169 publications

References 16 publications

Predicting missing links via significant paths

Predicting missing links via significant paths

A Survey on Link Prediction Problem in Social Networks

Information filtering in evolving online networks

Contact Info

Product

Resources

About