Cluster Nodes Research Articles

Multi-modal and multi-layer robustness analysis of the European rail and air networks

The robustness of long-distance transport services is paramount for ensuring network connectivity under disruptions. We conduct a comparative analysis of the European rail and air networks of 124 main metropolitan areas, assessing their ability to withstand successive network degradation. We undertake a multi-modal and multi-layer approach in our analysis of the robustness of long-distance transport networks. In particular, we are interested in the role of individual nodes for both air and rail networks as well as for the integrated multi-modal network. Given the hierarchical nature of long-distance transport services, we adopt a multi-layer perspective by means of clustering nodes based on their criticality in order to identify common performance profiles. Original metrics are formulated to measure the impact of nodes on network fragmentation, both at the individual level and as cluster members. Additionally, a new metric is introduced to assess cities’ reliance on air transportation, when considering the integrated multi-layer air-rail network during disruptions in air connections. Our findings indicate that the air network exhibits significantly greater robustness compared to rail, i.e. 10% versus 71% performance loss in the worst case scenario, respectively. Furthermore, primary sub-graph of nodes whose protection from attacks can greatly enhance network’s overall robustness is identified. We discuss our findings in terms of the relationship between network structure, robustness, and the role of critical nodes as well as propose potential mitigation measures.

Synthesis, structural characterization and third-order nonlinear optical study of W/Cu/S cluster-based coordination polymers with thioether-containing ligands

The assembly of [Et4N][Tp*WS3] (Et = ethyl, Tp* = tetrakis(3,5-dimethyl-1-pyrazolyl)borate), Cu(I) with 1,3,5-tri((pyridin-4-ylthio)methyl)benzene (tptmb) or 4,4′-bis((pyridin-4-ylthio)methyl)-1,1′-biphenyl (bptmb) leads to two coordination polymer with the formulae of {[Tp*WS3Cu3(tptmb)(CH3CN)](PF6)2·CHCl3·2CH3CN}n (1·CHCl3·2CH3CN) and {[(Tp*WS3Cu3)2(bptmb)3(CH3CN)](PF6)4·5CH2Cl2·1.5CH3CN}n (2·5CH2Cl2·1.5CH3CN), respectively. Compounds 1 and 2 were characterized by elemental analysis, infrared spectroscopy, UV–vis spectroscopy, high-resolution electrospray ionization mass spectrometry, and single-crystal X-ray diffraction. In 1, each tptmb ligand connects two adjacent cluster nodes by its three coordination arms, forming a cluster-based one-dimensional zigzag chain. In 2, equal amounts of [Tp*WS3Cu3(CH3CN)]2+ and [Tp*WS3Cu3]2+ clusters serve as three-connected nodes and are alternatively bridged by the bptmb ligands, forming a 2D network. Z-scan tests of 1 and 2 in dimethylformamide (DMF) solution reveal they have reverse saturable absorption with effective nonlinear absorption coefficients of 1.50 × 10−11 m/W for 1 and 5.00 × 10−11 m/W for 2. This work could inspire the construction of coordination polymer structures using cluster nodes and flexible thioether-containing ligands, providing a new way for the development of third-order nonlinear optical materials.

Disruptions in segregation mechanisms in fMRI-based brain functional network predict the major depressive disorder condition

This study investigates the functional brain network in major depressive disorder using network theory and a consensus network approach. At the macroscopic level, we found significant differences in connectivity measures such as node strength and clustering coefficient, with healthy controls exhibiting higher values. This is consistent with disruptions in functional brain network segregation in patients with major depressive disorder. Consensus network analysis revealed that the central executive and salience networks were predominant in healthy controls, whereas depressed patients showed greater overlap with the default mode network. No differences were found in network efficiency measures, indicating comparable brain network integration between healthy controls and major depressive disorder groups. Importantly, the clustering coefficient emerged as an effective diagnostic biomarker for depression, achieving high sensitivity (90%), specificity (92%), and overall precision (90%). Further analysis at the mesoscale level uncovered unique functional connections distinguishing healthy controls and major depressive disorder groups. Our findings underscore the utility of analyzing functional networks from the macroscale to the mesoscale, and provide insight into overcoming the challenges associated with intersubject variability and the multiple comparisons problem in network analysis.

Overlapping community detection in weighted networks via hierarchical clustering.

In real-world networks, community structures often appear as tightly connected clusters of nodes, with recent studies suggesting a hierarchical organization where larger groups subdivide into smaller ones across different levels. This hierarchical structure is particularly complex in trade networks, where actors typically belong to multiple communities due to diverse business relationships and contracts. To address this complexity, we present a novel algorithm for detecting hierarchical structures of overlapping communities in weighted networks, focusing on the interdependency between internal and external quality metrics for evaluating the detected communities. The proposed Graph Hierarchical Agglomerative Clustering (GHAC) approach utilizes maximal cliques as the basis units for hierarchical clustering. The algorithm measures dissimilarities between clusters using the minimal closed trail distance (CT-distance) and the size of maximal cliques within overlaps, capturing the density and connectivity of nodes. Through extensive experiments on synthetic networks with known ground truth, we demonstrate that the adjusted Silhouette index is the most reliable internal metric for determining the optimal cut in the dendrogram. Experimental results indicate that the GHAC method is competitive with widely used community detection techniques, particularly in networks with highly overlapping communities. The method effectively reveals the hierarchical structure of communities in weighted networks, as demonstrated by its application to the OECD weighted trade network, which describes the balanced trade value of bilateral trade relations.

Optimizing Multi-Tier Scheduling and Secure Routing in Edge-Assisted Software-Defined Wireless Sensor Network Environment Using Moving Target Defense and AI Techniques

Software Defined Wireless Sensor Networks (SDWSN) enable flexibility in Wireless Sensor Network (WSN) environments by defining the controllable functions to WSN nodes by the Software Defined Network (SDN) controller. Due to the rapid evolution of SDWSNs, adverse effects also have occurred in terms of interference, energy consumption, and security issues. Several state-of-the-art works lend their utmost best to the SDWSN environment. However, the complete picture (i.e., relatability and security in SDWSN) poses severe challenges. The state-of-the-art issues is addressed in this research by proposing interference-aware Multi-Tier Scheduling for the SDWSN environment (MTS-SDWSN). First, we perform network construction in which the proposed network is constructed in a 2D hexagonal grid structure to resolve the connectivity issue. Upon constructing the network, the SDWSN nodes are clustered and managed to reduce the energy consumption using the Divide Well To Merge Better (DWTMB) algorithm in which the optimal Cluster Leader (CL) is selected based on adequate constraint. The data from the clustered nodes are sent to the Local Base Station (LBS) via CL in which they are scheduled in multi-tier format to diminish the complexity and interference issues. The first tier involved in scheduling among Cluster Members (CMs) and CL using adequate metrics, whereas the successive tiers (i.e., second and third) involved in scheduling among CLs to LBSs and LBSs to Sink Node (SN) are done using the Non-Cooperative Fuzzy Theory (NCFT) method. Last, the scheduled nodes are routed to appropriate destinations using Secure and Optimal Routing Protocol (SORP). The proposed SORP includes the Alibaba and Forty Thieves (AFT) and Multi Criteria Decision Making (MCDM) algorithms for selecting and ranking the optimal routes. Further, the security of the routes is enabled by adopting trust and Moving Target Defense (MTD) mechanisms. The MTD includes route switching among the SDWSN devices and active switch handling using Cycle Generative Adversarial Networks (CGAN) among the switches. The proposed work is implemented using a NS-3.26 simulation tool, and performance of the proposed model and existing works shows that the proposed work outperforms the existing works.

Identifying successful football teams in the European player transfer network

This paper considers the European transfer market for professional football players as a network to study the relation between a team’s position in this network and performance in its domestic league. Our analysis is centered on eight top European leagues. The market in each season is represented as a weighted directed network capturing the transfers of players to or from the teams in these leagues, and we also consider the cumulative network over the past 28 years. We find that the overall structure of this transfer market network has properties commonly observed in real-world networks, such as a skewed degree distribution, high clustering, and small-world characteristics. To assess football teams we first construct a measure of within-league performance that is comparable across leagues. Regression analysis is used to relate league performance with both the network position and level of engagement of the team in the transfer market, under two complimentary setups. Network position variables include, e.g., betweenness centrality, closeness centrality and node clustering coefficient, whereas market engagement variables capture a team’s activity in the transfer market, e.g., total number of player transfers and total paid for players. For the season snapshots, the number of transfers correspond to weighted in- and out-degree. Our analysis first corroborates several recent findings relating aspects of market engagement with teams’ league performance. A higher number of incoming transfers indicates worse performance and better resourced teams perform better. Then, and across specifications, we find that network position variables remain salient even when engagement variables are already considered. This substantiates the notion in the existing literature that a high degree corresponds to better team performance and suggests that network aspects of trading strategy may affect a team’s success in their respective domestic league (or vice versa). In this sense, the approach and findings presented in this paper may in the future guide team’s player acquisition policies.

Cluster-efficient pangenome graph construction with nf-core/pangenome.

Pangenome graphs offer a comprehensive way of capturing genomic variability across multiple genomes. However, current construction methods often introduce biases, excluding complex sequences or relying on references. The PanGenome Graph Builder (PGGB) addresses these issues. To date, though, there is no state-of-the-art pipeline allowing for easy deployment, efficient and dynamic use of available resources, and scalable usage at the same time. To overcome these limitations, we present nf-core/pangenome, a reference-unbiased approach implemented in Nextflow following nf-core's best practices. Leveraging biocontainers ensures portability and seamless deployment in HPC environments. Unlike PGGB, nf-core/pangenome distributes alignments across cluster nodes, enabling scalability. Demonstrating its efficiency, we constructed pangenome graphs for 1000 human chromosome 19 haplotypes and 2146 E. coli sequences, achieving a two to threefold speedup compared to PGGB without increasing greenhouse gas emissions. Nf-core/pangenome is released under the MIT open-source license, available on GitHub and Zenodo, with documentation accessible at https://nf-co.re/pangenome/1.1.2/docs/usage. Supplementary data are available at Bioinformatics online.

GSSCL: A framework for Graph Self-Supervised Curriculum Learning based on clustering label smoothing

Graph self-supervised learning is an effective technique for learning common knowledge from unlabeled graph data through pretext tasks. To capture the interrelationships between nodes and their essential roles globally, existing methods use clustering labels as self-supervised signals. However, in some cases, these methods may introduce noise, leading to over-fitting of the model and a reduction in performance. To address these issues, a novel framework for Graph Self-Supervised Curriculum Learning based on clustering label smoothing called GSSCL has been proposed. GSSCL clusters knowledge in an easy-to-difficult manner, reducing the heavy dependence on the reliability of clustering and improving the generalizability of the model. Moreover, the Silhouette Coefficient is employed to evaluate the clustering confident scores for all nodes. Some nodes are selected based on high confident scores to perform self-supervised learning. To account for the possibility of complex heterophilous information in graphs (e.g., noisy links), clustering pseudo-label smoothing is performed on K-nearest neighbor graphs built upon the similarities between node features instead of the original graph structures. The obtained multi-scale knowledge is then applied to curriculum learning. Finally, comprehensive experiments conducted across diverse public graph benchmarks demonstrate the superior performance of the proposed framework. It exhibits comparable results to state-of-the-art methods across semi-supervised node classification and clustering tasks.

TCα-PIA: A Personalized Social Network Anonymity Scheme via Tree Clustering and α-Partial Isomorphism

Social networks have become integral to daily life, allowing users to connect and share information. The efficient analysis of social networks benefits fields such as epidemiology, information dissemination, marketing, and sentiment analysis. However, the direct publishing of social networks is vulnerable to privacy attacks such as typical 1-neighborhood attacks. This attack can infer the sensitive information of private users using users’ relationships and identities. To defend against these attacks, the k-anonymity scheme is a widely used method for protecting user privacy by ensuring that each user is indistinguishable from at least k−1 other users. However, this approach requires extensive modifications that compromise the utility of the anonymized graph. In addition, it applies uniform privacy protection, ignoring users’ different privacy preferences. To address the above challenges, this paper proposes an anonymity scheme called TCα-PIA (Tree Clustering and α-Partial Isomorphism Anonymization). Specifically, TCα-PIA first constructs a similarity tree to capture subgraph feature information at different levels using a novel clustering method. Then, it extracts the different privacy requirements of each user based on the node cluster. Using the privacy requirements, it employs an α-partial isomorphism-based graph structure anonymization method to achieve personalized privacy requirements for each user. Extensive experiments on four public datasets show that TCα-PIA outperforms other alternatives in balancing graph privacy and utility.

7914 Staged Treatment with Radioiodine Following Lobectomy in a Case of High Risk Papillary Thyroid Carcinoma

Abstract Disclosure: L. Pathak: None. A. Sharma: None. K. Rao: None. M.G. Jakoby: None. Introduction: Radioiodine ([1]3[1]I ) is usually reserved for patients with differentiated thyroid carcinoma (DTC) who are at high risk of persistence or recurrence after thyroidectomy. We present a case of papillary thyroid carcinoma (PTC) with extrathyroidal extension managed by lobectomy and staged [1]3[1]I ablation of the contralateral lobe and metastases due to comorbidities that made oncologic resection prohibitive. Case Description: A 58-year-old male had a 2.2 cm focus of PTC in the left thyroid lobe that extended into extrathyroidal soft tissue, adjacent trachea, and completely compromised the left recurrent laryngeal nerve. Due to left vocal cord paralysis and subglottic tracheal stenosis, left lobectomy was performed but oncologic resection (total thyroidectomy and tracheal resection) was deferred. The patient underwent recombinant human TSH (rhTSH) stimulated treatment with 30 mCi (1.1 GBq) of [1]3[1]I for right thyroid lobe ablation followed 60 d later by 180 mCi (6.7 GBq) of [1]3[1]I to treat metastases. Response to treatment was incomplete, with unstimulated thyroglobulin level falling from 293.9 to 18.9 ng/mL 1 y after [1]3[1]I and CT of the neck notable for a 2.7 cm left thyroid bed mass with invasion of adjacent trachea and a 1.0 cm right paratracheal mass, both of which exhibited hypermetabolism on PET imaging and yielded cytology consistent with poorly differentiated thyroid carcinoma (PDTC). A cluster of left supraclavicular lymph nodes showed both hypermetabolism on PET and radioiodine uptake on rhTSH scan. This patient is being prepared for surgical resection of the left thyroid bed and right paratracheal tumors and left lateral lymph node dissection. Discussion: Radioiodine lobar ablation is an alternative to completion thyroidectomy in patients with an unexpected diagnosis of DTC after lobectomy, with a recent meta-analysis finding a success rate of 69%. No difference in thyroid cancer recurrence rates was found in nearly 500 matched pairs of patients undergoing completion thyroidectomy or radioiodine lobar ablation. Unfortunately, staged lobar ablation followed by high dose [1]3[1]I was not successful in treating this high-risk PTC patient. Cases of coexisting PTC and PDTC are reported, and [1]3[1]I treatment was unsuccessful at treating PDTC that was not apparent on postsurgical histopathology. This case indicates that surgery is preferable to staged [1]3[1]I treatment of patients with aggressive PTC, especially if there are no comorbidities that make total thyroidectomy a high-risk procedure. Presentation: 6/2/2024

Comprehensive Evaluation and Selection of Cardamom (Elettaria cardamomum (L.) Maton) Germplasm Using Morphological Traits.

Cardamom (Elettaria cardamomum (L.) Maton) plays a crucial role in Guatemala's agriculture, supporting local families and covering 169,429.29 ha (making it the world's leading producer). Since its introduction to Guatemala in 1910, limited research has focused on unraveling the diversity and defining morphological traits critical for selecting excellent accessions. In this study, we examined 17 morphological traits across 288 accessions to identify key features associated with the germplasm. The comprehensive analysis employed principal component analysis, a morphological composite value (F-value), linear regression, and hierarchical clustering. The Shannon-Wiener diversity index ranged from 0.10 to 2.02, indicating the variation in diversity among traits. Principal component analysis and hierarchical clustering revealed six distinct germplasm groups. The comprehensive analysis facilitated the selection of 14 excellent accessions, and the regression equation incorporating criteria such as plant height, capsule color, panicle number per plant, panicle length, rhizome color, cluster number per panicle, cluster node length, and capsule number per cluster to identify cardamom germplasm. To develop a conservation strategy for the two putative foreign varieties ('Malabar' and 'Mysore'/'Vazhukka') introduced in Guatemala based on plant height, another 12 accessions were selected with a second comprehensive evaluation. This information offers insights into cardamom diversity for informed selection enhancing national utilization, productivity, and conservation.

Multi-level discriminator based contrastive learning for multiplex networks

Graph embedding is a technique for obtaining low-dimensional representations of nodes across diverse networks, which may then be used for various downstream tasks and applications. When it applies to heterogeneous networks, it is hard to handle heterogeneous networks because they usually contain different types of nodes and edges with more semantic and structural information. Recently, contrastive learning has developed as the preferred strategy for dealing with unsupervised heterogeneous graph embedding to reduce the cost of human label annotation. However, most multi-view contrastive learning approaches calculate the model’s loss only based on the mutual dependence between the node representation and graph representation. These approaches ignore that both node attributes and node clustering contain discriminative content. To solve this issue, we propose a model called Multi-Level Discriminator-based Contrastive Learning for Multiplex Networks (MLDCL). This model adopts a multi-level multi-discriminator-based approach that can simultaneously learn the global-level structural information, node-level attribute information, and local-level clustering information. Moreover, an augmentation strategy in the contrast learning process from the spectral domain is proposed to improve the representation and discriminative ability of MLDCL. Numerous tests with node clustering and classification tasks on widely used datasets demonstrate the efficacy of the proposed approach.

Graph-Structure-Based Multigranular Belief Fusion for Human Activity Recognition.

The belief functions (BFs) introduced by Shafer in the mid of 1970s are widely applied in information fusion to model epistemic uncertainty and to reason about uncertainty. Their success in applications is however limited because of their high-computational complexity in the fusion process, especially when the number of focal elements is large. To reduce the complexity of reasoning with BFs, we can envisage as a first method to reduce the number of focal elements involved in the fusion process to convert the original basic belief assignments (BBAs) into simpler ones, or as a second method to use a simple rule of combination with potentially a loss of the specificity and pertinence of the fusion result, or to apply both methods jointly. In this article, we focus on the first method and propose a new BBA granulation method inspired by the community clustering of nodes in graph networks. This article studies a novel efficient multigranular belief fusion (MGBF) method. Specifically, focal elements are regarded as nodes in the graph structure, and the distance between nodes will be used to discover the local community relationship of focal elements. Afterward, the nodes belonging to the decision-making community are specially selected, and then the derived multigranular sources of evidence can be efficiently combined. To evaluate the effectiveness of the proposed graph-based MGBF, we further apply this new approach to combine the outputs of convolutional neural networks + attention (CNN + Attention) in the human activity recognition (HAR) problem. The experimental results obtained with real datasets prove the potential interest and feasibility of our proposed strategy with respect to classical BF fusion methods.

Research on brain functional network property analysis and recognition methods targeting brain fatigue

At present, researches on brain fatigue recognition are still in the stage of single task and simple brain region network features, while researches on high-order brain functional network features and brain region state mechanisms during fatigue in multi-task scenarios are still insufficient, making it difficult to meet the needs of fatigue recognition under complex conditions. Therefore, this study utilized functional near-infrared spectroscopy (fNIRS) technology to explore the correlation and differences in the low-order and high-order brain functional network attributes of three task induced mental fatigue, and to explore the brain regions that have a major impact on mental fatigue. Self-training algorithms were used to identify the three levels of brain fatigue. The results showed that during the fatigue development, the overall connection strength of the endothelial cell metabolic activity and neural activity frequency bands of the low-order brain functional network first decreased and then increased, while the myogenic activity and heart rate activity frequency bands showed the opposite pattern. Network topology analysis indicated that from no fatigue to mild fatigue, the clustering coefficient of endothelial cell metabolic activity and myogenic activity frequency bands significantly decreased, while the characteristic path length of myogenic activity significantly increased; when experiencing severe fatigue, the small-world attribute of the neural frequency band significantly weakened. However, each frequency band maintained its small-world attribute, reflecting the self-optimization and adaptability of the network during the fatigue process. During mild fatigue, neuronal activity bands’ node degree, cluster coefficient, and efficiency rose in high-order brain networks, while low-order networks showed no significant changes. As fatigue progressed, the myogenic activity bands of high-order network properties dominated, but neural bands gained prominence in mild fatigue, approaching the level of myogenic bands in severe fatigue, indicating that brain fatigue orchestrated a shift from myogenic to neural dominance in frequency bands. In addition, during the process of fatigue, the four network attributes of the high-order network cluster composed of low-order nodes related to the prefrontal cortex region, left anterior motor region, motor assist region, and left frontal lobe eye movement region significantly increased, indicating that these brain regions had a significant impact on brain fatigue status. The accuracy of using both high-order and low-order features to identify fatigue levels reached 88.095%, indicating that the combined network features of both high-order and low-order fNIRS signals could effectively detect multi-level mental fatigue, providing innovative ideas for fatigue warning.

Systematic bibliometric and visualized analysis of research hotspots and trends on the application of artificial intelligence in glaucoma from 2013 to 2022.

To conduct a bibliometric analysis of research on artificial intelligence (AI) in the field of glaucoma to gain a comprehensive understanding of the current state of research and identify potential new directions for future studies. Relevant articles on the application of AI in the field of glaucoma from the Web of Science Core Collection were retrieved, covering the period from January 1, 2013, to December 31, 2022. In order to assess the contributions and co-occurrence relationships among different countries/regions, institutions, authors, and journals, CiteSpace and VOSviewer software were employed and the research hotspots and future trends within the field were identified. A total of 750 English articles published between 2013 and 2022 were collected, and the number of publications exhibited an overall increasing trend. The majority of the articles were from China, followed by the United States and India. National University of Singapore, Chinese Academy of Sciences, and Sun Yat-sen University made significant contributions to the published works. Weinreb RN and Fu HZ ranked first among authors and cited authors. American Journal of Ophthalmology is the most impactful academic journal in the field of AI application in glaucoma. The disciplinary scope of this field includes ophthalmology, computer science, mathematics, molecular biology, genetics, and other related disciplines. The clustering and identification of keyword nodes in the co-occurrence network reveal the evolving landscape of AI application in the field of glaucoma. Initially, the hot topics in this field were primarily "segmentation", "classification" and "diagnosis". However, in recent years, the focus has shifted to "deep learning", "convolutional neural network" and "artificial intelligence". With the rapid development of AI technology, scholars have shown increasing interest in its application in the field of glaucoma. Moreover, the application of AI in assisting treatment and predicting prognosis in glaucoma may become a future research hotspot. However, the reliability and interpretability of AI data remain pressing issues that require resolution.

Chemistry of Formate and Water Ligands on Metal Oxide Cluster Nodes of Metal-Organic Framework hcp Hf-UiO-66: Keys to Understanding Reactivity of Paired μ2-OH and Defect Sites.

Many metal-organic frameworks (MOFs) incorporate nodes that are metal oxide clusters, and ligands that have been observed on these nodes include formates, acetates, water, hydroxyl groups, and others, all of which are potentially important in affecting reactivities for applications in separations, catalysis, and sensing. Formate is a common node ligand, arising from formic acid used as a modulator and from N,N-dimethylformamide used as a solvent in MOF syntheses. Yet only little work has been reported characterizing the reactivities of node formate ligands. Infrared spectra reported here show that formate bonds to two types of sites on the paired Hf6O8 nodes of hcp UiO-66, namely, defect and μ2-OH sites. Quantifying the number of formate ligands by 1H NMR spectroscopy of digested samples showed an almost equal number of formate ligands on the two sites, indicating the likelihood that they neighbor each other. These formate ligands interact with water molecules, reversibly switching their bonding from bidentate to monodentate. The formates on μ2-OH sites of hcp Hf-UiO-66 interact much more strongly with water than those on defect sites of the same node, and both interact more strongly than isolated defect sites of Hf-UiO-66. Correspondingly, the catalytic activities of hcp UiO-66 determined as turnover frequencies on each site are approximately twofold higher than those on UiO-66, bolstering the inference that methanol dehydration is catalyzed by a node defect site and a neighboring node μ2-OH site. The results show how MOFs, with their well-defined node structures, provide unprecedented opportunities to understand details of reactivities and catalysis on metal oxide clusters, in contrast to bulk metal oxide surfaces.

Model‐based clustering in simple hypergraphs through a stochastic blockmodel

AbstractWe propose a model to address the overlooked problem of node clustering in simple hypergraphs. Simple hypergraphs are suitable when a node may not appear multiple times in the same hyperedge, such as in co‐authorship datasets. Our model generalizes the stochastic blockmodel for graphs and assumes the existence of latent node groups and hyperedges are conditionally independent given these groups. We first establish the generic identifiability of the model parameters. We then develop a variational approximation Expectation‐Maximization algorithm for parameter inference and node clustering, and derive a statistical criterion for model selection. To illustrate the performance of our R package HyperSBM, we compare it with other node clustering methods using synthetic data generated from the model, as well as from a line clustering experiment and a co‐authorship dataset.

A versatile framework for attributed network clustering via K-nearest neighbor augmentation

AbstractAttributed networks containing entity-specific information in node attributes are ubiquitous in modeling social networks, e-commerce, bioinformatics, etc. Their inherent network topology ranges from simple graphs to hypergraphs with high-order interactions and multiplex graphs with separate layers. An important graph mining task is node clustering, aiming to partition the nodes of an attributed network into k disjoint clusters such that intra-cluster nodes are closely connected and share similar attributes, while inter-cluster nodes are far apart and dissimilar. It is highly challenging to capture multi-hop connections via nodes or attributes for effective clustering on multiple types of attributed networks. In this paper, we first present as an efficient approach to attributed hypergraph clustering (AHC). includes a carefully-crafted K-nearest neighbor augmentation strategy for the optimized exploitation of attribute information on hypergraphs, a joint hypergraph random walk model to devise an effective AHC objective, and an efficient solver with speedup techniques for the objective optimization. The proposed techniques are extensible to various types of attributed networks, and thus, we develop as a versatile attributed network clustering framework, capable of attributed graph clustering, attributed multiplex graph clustering, and AHC. Moreover, we devise with algorithmic designs tailored for GPU acceleration to boost efficiency. We have conducted extensive experiments to compare our methods with 19 competitors on 8 attributed hypergraphs, 16 competitors on 6 attributed graphs, and 16 competitors on 3 attributed multiplex graphs, all demonstrating the superb clustering quality and efficiency of our methods.

Structural link prediction model with multi-view text semantic feature extraction

The exponential expansion of information has made text feature extraction based on simple semantic information insufficient for the multidimensional recognition of textual data. In this study, we construct a text semantic structure graph based on various perspectives and introduce weight coefficients and node clustering coefficients of co-occurrence granularity to enhance the link prediction model, in order to comprehensively capture the structural information of the text. Firstly, we jointly build the semantic structure graph based on three proposed perspectives (i.e., scene semantics, text weight, and graph structure), and propose a candidate keyword set in conjunction with an information probability retrieval model. Subsequently, we propose weight coefficients of co-occurrence granularity and node clustering coefficients to improve the link prediction model based on the semantic structure graph, enabling a more comprehensive acquisition of textual structural information. Experimental results demonstrate that our research method can reveal potential correlations and obtain more complete semantic structure information, while the WPAA evaluation index validates the effectiveness of our model.

Collaborative graph neural networks for augmented graphs: A local-to-global perspective

In the field of graph neural networks (GNNs) for representation learning, a noteworthy highlight is the potential of embedding fusion architectures for augmented graphs. However, prevalent GNN embedding fusion architectures mainly focus on handling graph combinations from a global perspective, often ignoring their collaboration with the information of local graph combinations. This inherent limitation constrains the ability of the constructed models to handle multiple input graphs, particularly when dealing with noisy input graphs collected from error-prone sources or those resulting from deficiencies in graph augmentation methods. In this paper, we propose an effective and robust embedding fusion architecture from a local-to-global perspective termed collaborative graph neural networks for augmented graphs (LoGo-GNN). Essentially, LoGo-GNN leverages a pairwise graph combination scheme to generate local perspective inputs. Together with the global graph combination, this serves as the basis to generate a local-to-global perspective. Specifically, LoGo-GNN employs a perturbation augmentation strategy to generate multiple augmentation graphs, thereby facilitating collaboration and embedding fusion from a local-to-global perspective through the use of graph combinations. In addition, LoGo-GNN incorporates a novel loss function for learning complementary information between different perspectives. We also conduct theoretical analysis to assess its expressive power under ideal conditions, demonstrating the effectiveness of LoGo-GNN. Our experiments, focusing on node classification and clustering tasks, highlight the superior performance of LoGo-GNN compared to state-of-the-art methods. Additionally, robustness analysis further confirms its effectiveness in addressing uncertainty challenges.