Neighborhood and PageRank methods for pairwise link prediction

Nassar, Huda; Benson, Austin R.; Gleich, David F.

doi:10.1007/s13278-020-00671-6

Cited by 24 publications

(13 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Recently, higher-order patterns in networks have attracted much research interest because of their tremendous success in many real-world applications (Milo et al, 2002;Alon, 2007) including discovering data insights (Benson et al, 2016;Paranjape et al, 2017;Benson et al, 2018;Lambiotte et al, 2019;Do et al, 2020) and building scalable computing algorithms (Yin et al, 2017;Paranjape et al, 2017;Fu et al, 2020;Veldt et al, 2020). Previous works on higher-order structure prediction can be generally grouped into two categories, predicting multiple edges/subgraphs in graphs (Lahiri and Berger-Wolf, 2007;Meng et al, 2018;Nassar et al, 2020;Cotta et al, 2020) and predicting hyperedges in hypergraphs Benson et al, 2018;Yadati et al, 2020;Alsentzer et al, 2020). Subgraphs, e.g., cliques of nodes (Benson et al, 2016), could be used to describe higher-order patterns.…”

Section: Related Workmentioning

confidence: 99%

“…On the other hand, learning dynamics from temporal networks is also challenging. It is typically hard to incorporate simple heuristics such as commute time and PageRanks to encode graph structures to elaborate the complex patterns embedded in the temporal network (Liben-Nowell and Kleinberg, 2007;Benson et al, 2018;Rossi et al, 2020b;Nassar et al, 2020) due to their limited model expressivity. Therefore, more powerful NN-based models have been introduced to this domain but mostly are designed for normal temporal graphs instead of temporal hypergraphs.…”

Section: Related Workmentioning

confidence: 99%

“…Heuristic Methods: Given a triplet of interest ({u, v}, w, t), we first project all the hyperedges before t into a static hypergraph. Then we transform the hypergraph to a traditional graph and calculate the corresponding features, including 3-way Adamic Adar Index (3-AA), 3-way Jaccard Coefficient (3-JC), 3-way Preferential Attachment(3-PA), and the arithmetic mean of traditional AA, JC and PA. We have not adopted the PageRank scores generalized for higher-order pattern prediction proposed by Nassar et al (2020) because computing those scores in temporal hypergraphs may take much time since the hypergraph structures change over time. The next step is to encode these features.…”

Section: C4 Baselines and Details Of The Experiments Setupmentioning

confidence: 99%

“…To evaluate our method, we compare HIT with ten baseline methods, including both heuristic methods and NN-based encoders. The detailed description of them is left in the Appendix C.4.Heuristic methods are the metrics previously used for edge prediction(Liben-Nowell and Kleinberg, 2007) and recently got generalized for hyperedge prediction(Benson et al, 2018;Nassar et al, 2020), including 3-way Adamic Adar Index (3-AA), 3-way Jaccard Coefficient (3-JC), 3-way Preferential Attachment(3-PA), and the Arithmetic Mean of traditional AA, JC and PA.…”

mentioning

confidence: 99%

See 3 more Smart Citations

Neural Predicting Higher-order Patterns in Temporal Networks

Liu¹,

Ma²,

Li³

2021

Preprint

View full text Add to dashboard Cite

Dynamic systems that consist of a set of interacting elements can be abstracted as temporal networks. Recently, higher-order patterns that involve multiple interacting nodes have been found crucial to indicate domain-specific laws of different temporal networks. This posts us the challenge of designing more sophisticated hypergraph models for these higher-order patterns and the associated new learning algorithms. Here, we propose the first model, named HIT, for higher-order pattern prediction in temporal hypergraphs. Particularly, we focus on predicting three types of common but important interaction patterns involving three interacting elements in temporal networks, which could be extended to even higher-order patterns. HIT extracts the structural representation of a node triplet of interest on the temporal hypergraph and uses it to tell what type of, when, and why the interaction expansion could happen in this triplet. HIT could achieve significant improvement (averaged 20% AUC gain to identify the interaction type, uniformly more accurate time estimation) compared to both heuristic and other neural-network-based baselines on 5 real-world large temporal hypergraphs. Moreover, HIT provides a certain degree of interpretability by identifying the most discriminatory structural features on the temporal hypergraphs for predicting different higher-order patterns.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: C4 Baselines and Details Of The Experiments Setupmentioning

confidence: 99%

mentioning

confidence: 99%

See 2 more Smart Citations

Neural Predicting Higher-order Patterns in Temporal Networks

Liu¹,

Ma²,

Li³

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…A huge number of works are dedicated to motif counting, listing (also called enumeration), or checking for the existence of a given motif [22,30]. However, while a few recent schemes focus on predicting triangles [9,63,64], no works target the problem of general motif prediction, i.e., analyzing whether specified complex structures may appear in the data. As with link prediction, it would enable predicting the evolution of data, but also finding missing structures in the available data.…”

Section: Introductionmentioning

confidence: 99%

Motif Prediction with Graph Neural Networks

Besta¹,

Grob²,

Miglioli³

et al. 2021

Preprint

View full text Add to dashboard Cite

Link prediction is one of the central problems in graph mining. However, recent studies highlight the importance of higher-order network analysis, where complex structures called motifs are the first-class citizens. We first show that existing link prediction schemes fail to effectively predict motifs. To alleviate this, we establish a general motif prediction problem and we propose several heuristics that assess the chances for a specified motif to appear. To make the scores realistic, our heuristics consider -among others -correlations between links, i.e., the potential impact of some arriving links on the appearance of other links in a given motif. Finally, for highest accuracy, we develop a graph neural network (GNN) architecture for motif prediction. Our architecture offers vertex features and sampling schemes that capture the rich structural properties of motifs. While our heuristics are fast and do not need any training, GNNs ensure highest accuracy of predicting motifs, both for dense (e.g., k-cliques) and for sparse ones (e.g., k-stars). We consistently outperform the best available competitor by more than 10% on average and up to 32% in area under the curve. Importantly, the advantages of our approach over schemes based on uncorrelated link prediction increase with the increasing motif size and complexity. We also successfully apply our architecture for predicting more arbitrary clusters and communities, illustrating its potential for graph mining beyond motif analysis.

show abstract

Deep attention framework for retweet prediction enriched with causal inferences

Sun¹,

Liu²

2023

Appl Intell

View full text Add to dashboard Cite

Neighborhood and PageRank methods for pairwise link prediction

Cited by 24 publications

References 33 publications

Neural Predicting Higher-order Patterns in Temporal Networks

Neural Predicting Higher-order Patterns in Temporal Networks

Motif Prediction with Graph Neural Networks

Deep attention framework for retweet prediction enriched with causal inferences

Contact Info

Product

Resources

About