Identifying critical higher-order interactions in complex networks

Aktaş, Mehmet Emin; Nguyen, Thu D.; Jawaid, Sidra; Riza, Rakin; Akbaş, Esra

doi:10.1038/s41598-021-00017-y

Cited by 19 publications

(6 citation statements)

References 41 publications

(38 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It cannot accommodate the case where the causal effect of X on Y depends on the value of a factor Z (higher-order interaction), i.e., a conditional causal effect that depends on interaction between X and Z—for example, less cars (X) lead to more cycling (Y) only if the built environment accommodates cycling (Z). That means that even if a CLD with edge weights that reflect ‘the amount of causal impact’ made by each of the included causal effects could be formulated, this in combination with network analysis would still be a limited representation of reality as it only allows for pairwise and not for higher-order interactions 79 , 80 —while computational modelling approaches can account for higher-order interactions by combining a set of factors in one equation.…”

Section: Discussionmentioning

confidence: 99%

Using network analysis to identify leverage points based on causal loop diagrams leads to false inference

Crielaard,

Quax,

Sawyer

et al. 2023

Sci Rep

View full text Add to dashboard Cite

Network analysis is gaining momentum as an accepted practice to identify which factors in causal loop diagrams (CLDs)—mental models that graphically represent causal relationships between a system’s factors—are most likely to shift system-level behaviour, known as leverage points. This application of network analysis, employed to quantitatively identify leverage points without having to use computational modelling approaches that translate CLDs into sets of mathematical equations, has however not been duly reflected upon. We evaluate whether using commonly applied network analysis metrics to identify leverage points is justified, focusing on betweenness- and closeness centrality. First, we assess whether the metrics identify the same leverage points based on CLDs that represent the same system but differ in inferred causal structure—finding that they provide unreliable results. Second, we consider conflicts between assumptions underlying the metrics and CLDs. We recognise six conflicts suggesting that the metrics are not equipped to take key information captured in CLDs into account. In conclusion, using betweenness- and closeness centrality to identify leverage points based on CLDs is at best premature and at worst incorrect—possibly causing erroneous identification of leverage points. This is problematic as, in current practice, the results can inform policy recommendations. Other quantitative or qualitative approaches that better correspond with the system dynamics perspective must be explored.

show abstract

Section: Discussionmentioning

confidence: 99%

Using network analysis to identify leverage points based on causal loop diagrams leads to false inference

Crielaard,

Quax,

Sawyer

et al. 2023

Sci Rep

View full text Add to dashboard Cite

show abstract

“…However, this algorithm can only be applied to the public opinion hypernetwork model and has certain limitations. Aktas et al 19 proposed two Laplacian operators based on the diffusion model to identify important hyperedges in hypergraphs, and confirmed the performance advantages of the new method by comparing with general centrality indicators. In this paper, based on the hypergraph theory, combined with the minimum eigenvalue property of the Laplacian deleted matrix of the hypernetwork, a new indicator MEGL (Minimum Eigenvalue of Grounded Laplacian Matrix) is proposed to identify the important hyperedges in the hypernetwork.…”

Section: Introductionmentioning

confidence: 90%

Important hyperedge sorting method based on grounded Laplacian matrix

Bai

et al. 2023

International Conference on Computer Application and Information Security (ICCAIS 2022)

View full text Add to dashboard Cite

How to use the methods of quantitative analysis to evaluate the importance of nodes or groups in complex networks is one of the important issues to be solved urgently in the field of network science research. Compared with the complex network, in which an edge represents the direct adjacent relationship between two nodes, the unique "hyperedge" in a hypernetwork is more suitable for representing groups, teams, and community structures of multiple nodes. Identifying the importance of teams and communities in the network is more conducive to controlling information dissemination in groups accurately, suppressing community-based outbreaks, predicting team research results, and discovering important drug targets. Existing algorithms focus on identifying important nodes in hypernetworks, and mining for important hyperedges is rare. In this paper, based on the hypergraph theory, combined with the property of the minimum eigenvalue of the grounded Laplacian matrix of the hypernetwork, a new index MEGL is proposed to identify the important hyperedges in hypernetwork. And it is applied in the drug target hypernetwork, which can not only identify important targets, but also identify important drugs. This method has important guiding significance for our drug development and target prediction, and also has certain reference significance for identifying important teams and communities in the network.

show abstract

“…A hypergraph is a unique model of a graph with hyperedges. Unlike a regular graph where the degree of each edge is 2, hyperedge is degree-free; it can connect an arbitrary number of nodes [4,5,10]. Then, we develop a Seq-HyGAN consisting of a novel two-level attention-based neural network that generates the representations of sequences as hyperedges.…”

Section: Introductionmentioning

confidence: 99%

Seq-HyGAN: Sequence Classification via Hypergraph Attention Network

Saifuddin¹,

May²,

Tanvir³

et al. 2023

Preprint

View full text Add to dashboard Cite

Sequence classification has a wide range of real-world applications in different domains, such as genome classification in health and anomaly detection in business. However, the lack of explicit features in sequence data makes it difficult for machine learning models. While Neural Network (NN) models address this with learning features automatically, they are limited to capturing adjacent structural connections and ignore global, higher-order information between the sequences. To address these challenges in the sequence classification problems, we propose a novel Hypergraph Attention Network model, namely Seq-HyGAN. To capture the complex structural similarity between sequence data, we first create a hypergraph where the sequences are depicted as hyperedges and subsequences extracted from sequences are depicted as nodes. Additionally, we introduce an attention-based Hypergraph Neural Network model that utilizes a two-level attention mechanism. This model generates a sequence representation as a hyperedge while simultaneously learning the crucial subsequences for each sequence. We conduct extensive experiments on four data sets to assess and compare our model with several state-of-the-art methods. Experimental results demonstrate that our proposed Seq-HyGAN model can effectively classify sequence data and significantly outperform the baselines. We also conduct case studies to investigate the contribution of each module in Seq-HyGAN.

show abstract

Identifying critical higher-order interactions in complex networks

Cited by 19 publications

References 41 publications

Using network analysis to identify leverage points based on causal loop diagrams leads to false inference

Using network analysis to identify leverage points based on causal loop diagrams leads to false inference

Important hyperedge sorting method based on grounded Laplacian matrix

Seq-HyGAN: Sequence Classification via Hypergraph Attention Network

Contact Info

Product

Resources

About