Temporal Link Prediction Using Matrix and Tensor Factorizations

Dunlavy, Daniel M.; Kolda, Tamara G.; Acar, Evrim

doi:10.1145/1921632.1921636

Cited by 434 publications

(273 citation statements)

References 38 publications

Supporting

Mentioning

268

Contrasting

Unclassified

Order By: Relevance

“…In general, to measure the link prediction accuracy of a temporal network sequence the AUC was used successfully in the past ( [13], [14]). The AUC is a preferable measurement in the presence of imbalanced datasets such as social networks where link density is usually low.…”

Section: A Auc Measure Of Link Prediction Accuracymentioning

confidence: 99%

Imputation of missing links and attributes in longitudinal social surveys

Ouzienko

Obradović

2013

Mach Learn

View full text Add to dashboard Cite

Abstract-We propose a unified approach for imputation of the links and attributes in longitudinal social surveys which accounts for changing network topology and interdependence between the actor's links and attributes. The previous studies on the treatment of non-respondents in longitudinal social networks were mostly concerned with imputation of the missing links only or imputation effects on the networks statistics. For this study we conduct a set of experiments on synthetic and real life datasets with 20%-60% of nodes missing under four mechanisms. The obtained results were better than when using alternative methods which suggest that our method can be used as a viable imputation tool.

show abstract

Section: A Auc Measure Of Link Prediction Accuracymentioning

confidence: 99%

Imputation of missing links and attributes in longitudinal social surveys

Ouzienko

Obradović

2013

Mach Learn

View full text Add to dashboard Cite

show abstract

“…Recent work has started to model network dynamics in order to better predict link and structure formation over time [10,7], but this work focuses on unattributed graphs. Previous work in relational learning on attributed graphs either uses static network snapshots or significantly limits the amount of temporal information incorporated into the models.…”

Section: Related Workmentioning

confidence: 99%

Time-Evolving Relational Classification and Ensemble Methods

Rossi

Neville

2012

Advances in Knowledge Discovery and Data Mining

View full text Add to dashboard Cite

Abstract. Relational networks often evolve over time by the addition, deletion, and changing of links, nodes, and attributes. However, accurately incorporating the full range of temporal dependencies into relational learning algorithms remains a challenge. We propose a novel framework for discovering temporal-relational representations for classification. The framework considers transformations over all the evolving relational components (attributes, edges, and nodes) in order to accurately incorporate temporal dependencies into relational models. Additionally, we propose temporal ensemble methods and demonstrate their e↵ectiveness against traditional and relational ensembles on two real-world datasets. In all cases, the proposed temporal-relational models outperform competing models that ignore temporal information.

show abstract

“…Tucker decomposition [18] is a widely used tensor decomposition method. In temporal link prediction, a timeevolving network is represented by a third-order tensor, where one dimension corresponds to the time frames and the other two dimensions correspond to the node IDs [7]. For example, an email message is represented by…”

Section: Related Workmentioning

confidence: 99%

“…Machine learning techniques proposed for predicting unknown links use the known links in a graph as training data. While conventional techniques are based on the assumption that the links are static, recent work has attempted to predict temporal links in dynamic and time-evolving networks [2]- [7]. In this temporal link prediction, only links among nodes at the same time point are considered ( Fig.…”

Section: Introductionmentioning

confidence: 99%

Link Prediction Across Time via Cross-Temporal Locality Preserving Projections

Oyama

Hayashi

Kashima

2012

IEICE Trans. Inf. & Syst.

View full text Add to dashboard Cite

SUMMARYLink prediction is the task of inferring the existence or absence of certain relationships among data objects such as identity, interaction, and collaboration. Link prediction is found in various applications in the fields of information integration, recommender systems, bioinformatics, and social network analysis. The increasing interest in dynamically changing networks has led to growing interest in a more general link prediction problem called temporal link prediction in the data mining and machine learning communities. However, only links among nodes at the same time point are considered in temporal link prediction. We propose a new link prediction problem called cross-temporal link prediction in which the links among nodes at different time points are inferred. A typical example of cross-temporal link prediction is cross-temporal entity resolution to determine the identity of real entities represented by data objects observed in different time periods. In dynamic environments, the features of data change over time, making it difficult to identify cross-temporal links by directly comparing observed data. Other examples of cross-temporal links are asynchronous communications in social networks such as Facebook and Twitter, where a message is posted in reply to a previous message. We adopt a dimension reduction approach to cross-temporal link prediction; that is, data objects in different time frames are mapped into a common low-dimensional latent feature space, and the links are identified on the basis of the distance between the data objects. The proposed method uses different low-dimensional feature projections in different time frames, enabling it to adapt to changes in the latent features over time. Using multi-task learning, it jointly learns a set of feature projection matrices from the training data, given the assumption of temporal smoothness of the projections. The optimal solutions are obtained by solving a single generalized eigenvalue problem. Experiments using a real-world set of bibliographic data for cross-temporal entity resolution and a real-world set of emails for unobserved asynchronous communication inference showed that introducing time-dependent feature projections improved the accuracy of link prediction.

show abstract

Temporal Link Prediction Using Matrix and Tensor Factorizations

Cited by 434 publications

References 38 publications

Imputation of missing links and attributes in longitudinal social surveys

Imputation of missing links and attributes in longitudinal social surveys

Time-Evolving Relational Classification and Ensemble Methods

Link Prediction Across Time via Cross-Temporal Locality Preserving Projections

Contact Info

Product

Resources

About