Deep node clustering based on mutual information maximization

Molaei, Soheila; Bousejin, Nima Ghanbari; Zare, Hadi; Jalili, Mahdi

doi:10.1016/j.neucom.2021.03.020

Cited by 14 publications

(4 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Subsequently, the proposed framework lists maximal cliques from the constructed graph in different hops to learn useful representations, as used by [27,47]. The Bron-Kerbosch method [45] is used to obtain all maximal cliques in the graph.…”

Section: Cliquefluxnetmentioning

confidence: 99%

CliqueFluxNet: Unveiling EHR Insights with Stochastic Edge Fluxing and Maximal Clique Utilisation using Graph Neural Networks

Molaei,

Bousejin,

Ghosheh

et al. 2024

Preprint

Self Cite

View full text Add to dashboard Cite

Electronic Health Records (EHRs) play a crucial role in shaping predictive healthcare models, yet they encounter challenges such as significant data gaps and class imbalances. Traditional Graph Neural Networks (GNNs) approaches have limitations in fully leveraging neighbourhood data or demanding intensive computational regularisation. To address this challenge, we introduce CliqueFluxNet, a novel framework that innovatively constructs a patient similarity graph to maximise cliques, thereby highlighting strong inter-patient connections. At the heart of CliqueFluxNet lies its stochastic edge fluxing strategy — a dynamic process involving random edge addition and removal during training. This strategy aims to enhance the model's generalizability and mitigate overfitting. Our empirical analysis, conducted on MIMIC-III and eICU datasets, demonstrates significant progress in representation learning, particularly in scenarios with limited data availability. Qualitative assessments further underscore CliqueFluxNet’s effectiveness in extracting meaningful EHR representations, solidifying its potential for advancing GNN applications in healthcare analytics.

show abstract

Section: Cliquefluxnetmentioning

confidence: 99%

CliqueFluxNet: Unveiling EHR Insights with Stochastic Edge Fluxing and Maximal Clique Utilisation using Graph Neural Networks

Molaei,

Bousejin,

Ghosheh

et al. 2024

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…transportation networks [4], and citation networks [5]. Recently, many studies have emerged on learning representations to encode structural information of the graph [6,7,8,9,10]. These graph representation learning algorithms convert graph data into a low-dimensional space.…”

Section: Introductionmentioning

confidence: 99%

Graph representation learning based on deep generative gaussian mixture models

et al. 2023

Self Cite

View full text Add to dashboard Cite

“…Downstream machine learning tasks can then use these latent vectors as feature inputs [5,6]. For example, COOL [7], and GHNN [8] employ graph representations in their classification task, Modularity-aware VGAE [9], and GCN-LP [10] in their link prediction tasks, and SOLI [11] in its clustering task.…”

Section: Introductionmentioning

confidence: 99%