Search

The manufacturing era is going through an evolutionary phase from Industry 3.0 to Industry 4.0 (I4.0). In 2011, Germany initiated I4.0 during the Hanover fair in collaboration with industrialists, academicians and researchers. The benefits of implementing I4.0 are attracting the curiosity of practitioners. In current academic literature, there is little discussion related to analysing potential roadblocks of I4.0 implementation using mathematical modelling. This paper has identified roadblocks through an extensive literature review and validation has been carried out by gathering experts’ opinions. As per the findings, the identified roadblocks have been categorized into five sections: management roadblocks, operational roadblocks, human resource roadblocks, procedural and behavioural roadblocks. A case study of the automobile industry has been discussed. The data was collected from experts in the area of I4.0. A Graph Theory and Matrix Approach (GTMA) was applied to evaluate the relative intensity of the roadblocks, Comprehensive I4.0 Index (CII) and its range for each category of roadblock. Implications for industrialists, practitioners and academicians are provided.

Enhancement of reliability is critical in designing and planning of distribution systems that operate efficiently and with minimal interruption to customer loads, because they use various types of resources and technologies to meet energy demand such as distributed generation (DG). DG is expected to play an enhanced role in emerging power systems. One of the major issues in the planning stage for future distribution systems is the optimal design of multiple micro-grids (MMGs) in distribution systems. This paper describes a novel method for clustering a traditional passive distribution system into a cluster of Multiple Micro Grids (MMGs). Clustering a traditional distribution system into MMGs provides numerous merits for costumers and distribution system operators, including a local control strategy to minimise interaction between various MGs, local reactive power compensation to prevent fault propagation, minimise power losses, and ultimately improving system reliability. In each MG area, DGs and reactive power resources are deployed to reach the acceptable level of reliability. The proposed clustering procedure is based on a weighted graph partitioning approach, with the weights representing the apparent power of the lines. Various reliability tests were performed in order to determine the best location for DG penetration in the distribution system. The passive distribution system becomes an active system after DG penetration. As a result, planners must investigate the influence of DG on distribution system reliability. The aim of this paper is to investigate the impact of DG on distribution system reliability. The results on the test distribution system demonstrate the proposed method’s performance and effectiveness.

Graph theory allows us to gain better insight concerning changes in brain functional architecture associated with cognitive impairments in the early stages of multiple sclerosis (MS). In the present study, we employed a machine-learning system based on graph measures from functional networks constructed by cognitive task-related functional magnetic resonance imaging (fMRI) data. We used a predefined atlas to define the brain regions and Pearson’s correlation to describe the connectivity strength between the regions. Then, several graph metrics were extracted for each subject. After that, the most efficient subsets of features were selected through the Wilcoxon rank-sum test, and the linear support vector machine (SVM) classifier was employed to distinguish between MS and healthy subjects. The node degree, subgraph centrality, K-coreness, and PageRank centralities measured in the left fusiform, hippocampus, and parahippocampal gyri regions demonstrated an accuracy of 85% through the combination of all local measures. Two optimal global measures, modularity and small-worldness index, and individual betweenness centrality feature improved the identification of MS patients with a sensitivity of 81.25%. Our results indicated the potential of the proposed system to identify cognitive changes in early MS for diagnostic purposes.

An airport and its region have a mutual interaction. Thus, the airport's performance is crucial for its region. Moreover, airport's performance is influenced by air transport network. The measurement of the quality and capacity of the air transport network is multifaceted. The study assesses an airport's network in terms of its type, accessibility, and connectivity. It focusses on 12 Tier-II Indian cities having medium-hub international airports because of their huge growth potential and under-addressed assessment. Furthermore, it is in coherence with the current policies of boosting regional air connectivity. Specific analyses for Tier-II cities' airports is important as a comparison with the bigger metropolises will be skewed. This exploration puts special emphasis on their relationship with the Tier-I cities as the connectivity of Indian airports is majorly complementary than competitive. Measures used for assessment are concentration indices, graph theory-based connectivity and accessibility indices (shortest path length and quickest time length), and partial accessibility indicators (daily accessibility, potential indicator, location indicator, and network efficiency). Modifications and synthesis of partial indicators are also attempted. Annual data sets for the financial year 2017–18 and real-time data from 25–July 31, 2018 are used, and the analysis was carried out from April–August 2018. The study identifies future potential (hubs/feeders) among the Tier-II cities' airports and proposes to align policies accordingly. The results illustrate the need for more strong hubs in northern India and reveal the potential of Amritsar, Calicut, and Trichy of becoming future regional hubs.

Aim: The aim of the present study is to investigate the relationship between both functional connectivity and brain networks with cognitive decline, in patients with Parkinson's disease (PD). Introduction: PD phenotype is not limited to motor impairment but, rather, a wide range of non-motor disturbances can occur, with cognitive impairment being one of the most common. However, how the large-scale organization of brain activity differs in cognitively impaired patients, as opposed to cognitively preserved ones, remains poorly understood. Methods: Starting from source-reconstructed resting-state magnetoencephalography data, we applied the phase linearity measurement (PLM) to estimate functional connectivity, globally and between brain areas, in PD patients with and without cognitive impairment (respectively PD-CI and PD-NC), as compared with healthy subjects (HS). Further, using graph analysis, we characterized the alterations in brain network topology and related these, as well as the functional connectivity, to cognitive performance. Results: We found reduced global and nodal PLM in several temporal (fusiform gyrus, Heschl's gyrus, and inferior temporal gyrus), parietal (postcentral gyrus), and occipital (lingual gyrus) areas within the left hemisphere, in the gamma band, in PD-CI patients, as compared with PD-NC and HS. With regard to the global topological features, PD-CI patients, as compared with HS and PD-NC patients, showed differences in multi-frequencies bands (delta, alpha, gamma) in the Leaf fraction, Tree hierarchy (Th) (both higher in PD-CI), and Diameter (lower in PD-CI). Finally, we found statistically significant correlations between the Montreal Cognitive Assessment test and both the Diameter in delta band and the Th in the alpha band. Conclusion: Our work points to specific large-scale rearrangements that occur selectively in cognitively compromised PD patients and are correlated to cognitive impairment. Impact statement In this article, we want to test the hypothesis that the cognitive decline observed in Parkinson's disease (PD) patients may be related to specific changes of both functional connectivity and brain network topology. Specifically, starting from magnetoencephalography signals and by applying the phase linearity measurement (PLM), a connectivity metric that measures the synchronization between brain regions, we were able to highlight differences in the global and nodal PLM values in PD patients with cognitive impairment as compared with both cognitively unimpaired patients and healthy subjects. Further, using graph analysis, we analyzed alterations in brain network topology that were related to cognitive functioning.

Vibration signals always contain noise and irregularities which makes spectrum analysis difficult to extract high-level features. Recently, graph theory has been applied to spectrum analysis to improve the performance of feature extraction. By converting the raw data into graphs, hidden structural and topological information can be obtained. In this paper, a spatial-temporal graph-based feature extraction, called SuperGraph, for rotating machinery fault diagnosis is proposed. Specifically, graph theory-based spectrum analysis is used to construct the spatial-temporal graph. Then, the Laplacian matrix-based feature vector is extracted from the constructed spatial-temporal graph. By this means, the spatial-temporal graph is converted into the one-dimensional vector for further constructing SuperGraph, where each node of the SuperGraph represents a spatial-temporal graph and the SuperGraph is composed of many local graphs. In the local graph, only the same type of nodes are connected to form a fully connected graph. Thus, the task of graph classification can be transformed into classifying the nodes in the SuperGraph. After graph convolutional network is established for learning and obtaining deep features, the label of nodes is identified from a softmax model. Experiments are conducted on two benchmarking datasets and a practical experimental platform to verify effectiveness of the proposed fau

Band selection (BS) aims to choose a salient subset implied sufficient information from the numerous bands, which supplies a significantly efficient way to alleviate the barrier of dimensionality disaster for hyperspectral image classification (HSIC). This letter develops a semisupervised BS approach based on dual-constrained low-rank representation BS (DCLRR-BS) with two regularizations for HSIC. To be specific, a low-rank representation model is first proposed with super-pixel and imbalanced class-wise constraints, which are explicitly integrated to improve the performance of the band description. Next, the clusters are built adaptively based on graph theory in an unsupervised manner to rapid selection efficiency. A selection criterion is last designed to highlight the prominent band of each subset cluster to fulfill the BS procedure. Experimental results conducted on four types of classifiers with two real hyperspectral image (HSI) data sets demonstrate that the proposed DCLRR-BS method performs well in the imbalanced HSIC area.

The paper unravels the potential of graph theory in the heart murmur auscultation by constructing a complex network from the single murmur time-series signals. In this study, forty-eight murmur signals of mitral incompetence (MI) and healthy heart (NM) are subjected to complex network and wavelet analyses. For the complex network analysis, the correlation coefficient is fixed as 0.8 and the time series is segmented based on the time delay obtained from the autocorrelation function. The signals are classified based on graph features, reflecting the haemodynamic through the mitral valve, using unsupervised principal component analysis and supervised k-nearest neighbour classifier. The appearance of many frequency components in the murmur due to MI is understood from the wavelet analysis. The graphs of NM and MI are found to show two well-defined clusters. When the graph of NM shows a large number of uncorrelated nodes due to the absence of signal in the systolic region that of MI shows interconnected nodes. The improper closing of the mitral valve and the regurgitation of blood in MI results in sound signals in the systolic region, responsible for the increased number of edges compared to NM. The present economical and sensitive graph-based method opens up the plausibility of remote auscultation in primary health centres in the context of the outbreak of pandemic COVID 19.

Objective. Action intention understanding EEG signal classification is indispensable for investigating human-computer interactions and intention understanding mechanisms. Numerous investigations on classification tasks extract classification features by using graph theory metrics; however, the classification results are usually not good. Method. To effectively implement the task of action intention understanding EEG signal classification, we proposed a new feature extraction method by improving discriminative spatial patterns. Results. The whole frequency band and fusion band achieved satisfactory classification accuracies. Compared with other authors’ methods for action intention understanding EEG signal classification, the new method performs more satisfactorily in some aspects. Conclusions. The new feature extraction method not only effectively avoids complex values when solving the generalized eigenvalue problem but also perfectly realizes appreciable classification accuracies. Fusing the classification features of different frequency bands is a useful strategy for the classification task.

In recent times, security and safety are, at least, conducted in safety-sensitive or critical sectors. Nevertheless, both processes do not commonly analyze the impact of security risks on safety. Several scholars are focused on integrating safety and security risk assessments, using different methodologies and tools in critical infrastructures (CIs). Bayesian networks (BN) and graph theory (GT) have received much attention from academia and industries to incorporate security and safety features for different CI applications. Hence, this study aims to conduct a systematic literature review (SLR) for co-engineering safety and security using BN or GT. In this SLR, the preferred reporting items for systematic reviews and meta-analyses recommendations (PRISMA) are followed. Initially, 2295 records (acquired between 2011 and 2020) were identified for screening purposes. Later on, 240 articles were processed to check eligibility criteria. Overall, this study includes 64 papers, after examining the pre-defined criteria and guidelines. Further, the included studies were compared, regarding the number of required nodes for system development, applied data sources, research outcomes, threat actors, performance verification mechanisms, implementation scenarios, applicability and functionality, application sectors, advantages, and disadvantages for combining safety, and security measures, based on GT and BN. The findings of this SLR suggest that BN and GT are used widely for risk and failure management in several domains. The highly focused sectors include studies of the maritime industry (14%), vehicle transportation (13%), railway (13%), nuclear (6%), chemical industry (6%), gas and pipelines (5%), smart grid (5%), network security (5%), air transportation (3%), public sector (3%), and cyber-physical systems (3%). It is also observed that 80% of the included studies use BN models to incorporate safety and security concerns, whereas 15% and 5% for GT approaches and joint GT and BN methodologies, respectively. Additionally, 31% of identified studies verified that the developed approaches used real-time implementation, whereas simulation or preliminary analysis were presented for the remaining methods. Finally, the main research limitations, concluding remarks and future research directions, are presented

Sophisticated network‐based approaches such as structural connectomics may help to detect a biomarker of mild traumatic brain injury (mTBI) in children. This study compared the structural connectome of children with mTBI or mild orthopedic injury (OI) to that of typically developing (TD) children. Children aged 8–16.99 years with mTBI (n = 83) or OI (n = 37) were recruited from the emergency department and completed 3T diffusion MRI 2–20 days postinjury. TD children (n = 39) were recruited from the community and completed diffusion MRI. Graph theory metrics were calculated for the binarized average fractional anisotropy among 90 regions. Multivariable linear regression and linear mixed effects models were used to compare groups, with covariates age, hemisphere, and sex, correcting for multiple comparisons. The two injury groups did not differ on graph theory metrics, but both differed from TD children in global metrics (local network efficiency: TD > OI, mTBI, d = 0.49; clustering coefficient: TD < OI, mTBI, d = 0.49) and regional metrics for the fusiform gyrus (lower degree centrality and nodal efficiency: TD > OI, mTBI, d = 0.80 to 0.96; characteristic path length: TD < OI, mTBI, d = −0.75 to −0.90) and in the superior and middle orbital frontal gyrus, paracentral lobule, insula, and thalamus (clustering coefficient: TD > OI, mTBI, d = 0.66 to 0.68). Both mTBI and OI demonstrated reduced global and regional network efficiency and segregation as compared to TD children. Findings suggest a general effect of childhood injury that could reflect pre‐ and postinjury factors that can alter brain structure. An OI group provides a more conservative comparison group than TD children for structural neuroimaging research in pediatric mTBI., Children with mTBI and OI demonstrated reduced global and regional network efficiency and segregation as compared to TD children. Findings suggest a general effect of childhood injury that could reflect pre‐ and postinjury factors that can alter brain structure. An OI group provides a more conservative comparison group than TD children for pediatric mTBI research.

With the development of hand-held photographic devices, document images in unconstrained environments can be captured in high-speed and high-resolution. It will be more efficient to process the text information of multiple documents simultaneously. In this paper, we propose a multi-document detection approach. It can estimate the amount of documents and also detect their accurate locations via iteratively searching the four corners and their direction maps from individual document in the image. Even for slightly occluded documents, the proposed method can infer the hidden corner positions. The model is designed to jointly learn the corner categories, locations and their directions in attentional regions via two branches of the same sequential prediction process. The association score is calculated based on the them between two corner connections. The graph theory, considering corners as nodes and association scores as edges, is applied to get the quadrangle for each document in image. For evaluation, we collect a Multi-Doc data set which contains 2,200 document images in various natural scenes. We show that the baseline model trained on this collected data and bench-mark SmartDoc 2015 can detect both single and multiple documents accurately and effectively.

Background Nowadays we are observing an explosion of gene expression data with phenotypes. It enables us to accurately identify genes responsible for certain medical condition as well as classify them for drug target. Like any other phenotype data in medical domain, gene expression data with phenotypes also suffer from being a very underdetermined system. In a very large set of features but a very small sample size domain (e.g. DNA microarray, RNA-seq data, GWAS data, etc.), it is often reported that several contrasting feature subsets may yield near equally optimal results. This phenomenon is known as instability. Considering these facts, we have developed a robust and stable supervised gene selection algorithm to select a set of robust and stable genes having a better prediction ability from the gene expression datasets with phenotypes. Stability and robustness is ensured by class and instance level perturbations, respectively. Results We have performed rigorous experimental evaluations using 10 real gene expression microarray datasets with phenotypes. They reveal that our algorithm outperforms the state-of-the-art algorithms with respect to stability and classification accuracy. We have also performed biological enrichment analysis based on gene ontology-biological processes (GO-BP) terms, disease ontology (DO) terms, and biological pathways. Conclusions It is indisputable from the results of the performance evaluations that our proposed method is indeed an effective and efficient supervised gene selection algorithm.

Social media has had a strong presence in many people’s lives over the last decade. In addition, social media platforms have allowed people to share opinions, provide advice on numerous factors, including where to visit, as well as to stay connected and maintain friendships. The hospitality and tourism industry, however, can make effective use of these powerful tools for marketing purposes, collaboration and information sharing, and service offerings. Reviewing social media followers’ behaviors and interests offers a wealth of information and valuable data for a variety of tourism organizations. This case study focuses on an analysis of the social networks applied to the fortified town of Fredrikstad in Norway. The data used in this research study were collected from the Facebook site of the tourist authority. The results of this research project demonstrate the strengths of applying a social network analysis to a dataset, which can aid in the strategic direction of a tourism destination. The conversations of the greatest interest can successfully be identified as well as the growth of the online network. This paper adds knowledge to the literature through the application of a social network analysis regarding the success of a tourism destination and its future potential.

Previous deep learning methods usually required large‐scale annotated data, which is computationally exhaustive and unrealistic in certain scenarios. Therefore, few‐shot learning, where only a few annotated training images are available for training, has attracted increasing attention these days, showing huge potential in practical applications, such as portable equipment or security inspection, and so on. However, current few‐shot learning methods usually neglect the valuable semantic correlations between samples, thereby failing in extracting discriminating relations to achieve accurate predictive results. In this work, extending on a recent state‐of‐the‐art few‐shot learning method, transductive relation‐propagation network (TRPN), which considers the correlations between training samples, a constrained relation‐propagation network is proposed to further regularise the distilled correlations and thus achieve favourable few‐shot classification performance. The proposed framework contains three main components, namely preprocess module, relational propagation module, and relation constraint module. First, sample features are extracted and a relation graph node is constructed by treating the relation of each support–query pair as a graph node in the preprocess module. After that, in the relation propagation module (RPM), the valuable information of support–query pairs is modelled and propagated to directly generate the relational representations for further prediction. Then, a relation constraint module is introduced to regularise the relational representations and make it consistent with the ground‐truth relations as much as possible. With the guidance of the effective RPM and relation constraint module, the relational representations of the support–query pairs are distinguishable and thus can achieve accurate predictive results. Comprehensive experiments conducted on widely used benchmarks validate the effectiveness of our method compared to state‐of‐the‐art few‐shot classification approaches.

The connections in a graph generate a structure that is independent of a coordinate system. This visual metaphor allows creating a more flexible representation of data than a two-dimensional scatterplot. In this work, we present STAD (Simplified Topological Abstraction of Data), a parameter-free dimensionality reduction method that projects high-dimensional data into a graph. STAD generates an abstract representation of high-dimensional data by giving each data point a location in a graph which preserves the approximate distances in the original high-dimensional space. The STAD graph is built upon the Minimum Spanning Tree (MST) to which new edges are added until the correlation between the distances from the graph and the original dataset is maximized. Additionally, STAD supports the inclusion of additional functions to focus the exploration and allow the analysis of data from new perspectives, emphasizing traits in data which otherwise would remain hidden. We demonstrate the effectiveness of our method by applying it to two real-world datasets: traffic density in Barcelona and temporal measurements of air quality in Castile and Leon in ´ Spain. The authors wish to thank Danai Kafetzaki for valuable feedback and proofreading. This work was supported in part by the IWT/SBO 150056 project ”ACquiring CrUcial Medical information Using LAnguage TEchnology” (ACCUMULATE), and by the Flanders AI Impulse Program (“Onderzoeksprogramma Artificiele Intelligentie (AI) ¨ Vlaanderen”).

Smart homes of the future will have to deal with multi-occupancy scenarios. Multi-occupancy systems entail a preliminary and critical feature: the capability of counting people. This can be fulfilled by means of simple binary sensors, cheaper and more privacy preserving than other sensors, such as cameras. However, it is currently unclear how many people can be counted in a smart home, given the set of available sensors. In this paper, we propose a graph-based technique that allows to map a smart home to an undirected graph G and discover the lower-bound of certainly countable people, also defined as certain count. We prove that every independent set of n vertices of an undirected graph G represents a minimum count of n people. We also prove that the maximum number of certainly countable people corresponds to the maximum independent sets of G, and that the maximal independent sets of G provide every combination of active sensors that ensure different minimum count. Last, we show how to use this technique to identify and optimise suboptimal deployment of sensors, so that the assumptions can be tightened and the theoretical lower-bound improved., Funding agency:Örebro University

It is common practice that central banks around the world must adopt an inflation targeting framework; based on the assumption that inflation could be reduced by increasing interest rates. On contrary, the theoretical literature and data-based evidences differ remarkably. The arduous attached in finding the true association and causation is the existence of multiple monetary transmission channels. Theoretical literature lists both monetary and non-monetary channels linking interest rate and inflation. However, most of the existing studies are focused on single equation model ignoring other parallel channels. This study is first of its kind where we have developed modified Peter and Clark (PC) algorithm of the Graph-Theoretic approach taking all the monetary and non-monetary channels to determine the causal nexus between monetary policy and inflation. The results show that, causality is running from interest rate towards inflation; suggesting a positive and significant long run relationship of interest rate with inflation in case of Pakistan. Furthermore, monetary policy have cost-side effects on inflation; however, the monetary policy becomes counterproductive whenever high interest rate is used to decrease cost push inflation. Therefore, there is need of serious rethinking about current monetary policy regime. JEL Codes - E4; E5; E42; E52

Industries are economically linked through customer-supplier trade flows. We show theoretically and empirically that industry shocks propagating along this inter-sectoral trade network can feed back to the originating industry, causing an "echo" -- intermediate-term autocorrelation in returns. Adopting techniques from graph theory, we find that the strength of the trade network feedback is a crucial determinant of the echo effect in industry returns. Returns of the echo strategy implemented within high-feedback strength industries exceed 1% monthly. Consistent with limited-information models, the relation between feedback strength and echo profits is strongest in industries with information diffusion frictions, such as low analyst coverage, along the feedback loop. Overall, our results identify inter-sectoral trade networks as important conduits of industry shocks and provide the first explanation for the echo effect.

Many materials with remarkable properties are structured as percolating nanoscale networks (PNNs). The design of this rapidly expanding family of composites and nanoporous materials requires a unifying approach for their structural description. However, their complex aperiodic architectures are difficult to describe using traditional methods that are tailored for crystals. Another problem is the lack of computational tools that enable one to capture and enumerate the patterns of stochastically branching fibrils that are typical for these composites. Here, we describe a computational package, StructuralGT, to automatically produce a graph theoretical (GT) description of PNNs from various micrographs that addresses both challenges. Using nanoscale networks formed by aramid nanofibers as examples, we demonstrate rapid structural analysis of PNNs with 13 GT parameters. Unlike qualitative assessments of physical features employed previously, StructuralGT allows researchers to quantitatively describe the complex structural attributes of percolating networks enumerating the network's morphology, connectivity, and transfer patterns. The accurate conversion and analysis of micrographs was obtained for various levels of noise, contrast, focus, and magnification, while a graphical user interface provides accessibility. In perspective, the calculated GT parameters can be correlated to specific material properties of PNNs (e.g., ion transport, conductivity, stiffness) and utilized by machine learning tools for effectual materials design.

This study aims to introduce Suzuki type Σcontraction mappings with simulation functions in the frame of modular b-metric spaces. Also, some coincidence and common fixed point results are obtained for four mappings using the weakly compatibility property that these results are the extensions and improvements of the existing literature. Finally, we also present two applications on graph theory and homotopy theory, which show applicability and validity of our results.

PurposeThe present study aims at developing a model to quantify supply chain resilience as a single numerical value. The numerical value is called resilience index that measures the resilience capability of the case company's supply chain. The model calculates the index value based on the interactions between the enablers of supply chain resilience and its dimensions.Design/methodology/approachGraph theoretic approach (GTA) is used to evaluate the resilience index for the case company's supply chain. In GTA, the dimensions of resilience enablers and their interdependencies are modelled through a digraph. The digraph depicting the influence of each dimension is converted into an adjacency matrix. The permanent function value of the adjacency matrix is called the resilience index (RI).FindingsThe proposed approach has been illustrated in context of an Indian automobile organization, and value of the RI is evaluated. The best case and the worst-case values are also obtained with the help of GTA. It is noted from the model that strategic level dimension of enablers is most important in contributing towards supply chain resilience. They are followed by tactical and operational level enablers. The GTA framework proposed will help supply chain practitioners to evaluate and benchmark the supply chain resilience of their respective organizations with the best in the industry.Originality/valueA firm can compare the RI of its own supply chain with other's supply chain or with the best in the industry for benchmarking purpose. Benchmarking of resilience will help organizations in developing strategies to compete in dynamic market scenario.

Epileptic EEG signal recognition is an important method for epilepsy detection. In essence, epileptic EEG signal recognition is a typical imbalanced classification task. However, traditional machine learning methods used for imbalanced epileptic EEG signal recognition face many challenges: (1) traditional machine learning methods often ignore the imbalance of epileptic EEG signals, which leads to misclassification of positive samples and may cause serious consequences and (2) the existing imbalanced classification methods ignore the interrelationship between samples, resulting in poor classification performance. To overcome these challenges, a graph-based extreme learning machine method (G-ELM) is proposed for imbalanced epileptic EEG signal recognition. The proposed method uses graph theory to construct a relationship graph of samples according to data distribution. Then, a model combining the relationship graph and ELM is constructed; it inherits the rapid learning and good generalization capabilities of ELM and improves the classification performance. Experiments on a real imbalanced epileptic EEG dataset demonstrated the effectiveness and applicability of the proposed method.

Several Alzheimer’s disease (AD) atrophy subtypes were identified, but their brain network properties are unclear. We analyzed data from two independent datasets, including 166 participants (103 AD/63 controls) from the DZNE-longitudinal cognitive impairment and dementia study and 151 participants (121 AD/30 controls) from the AD neuroimaging initiative cohorts, aiming to identify differences between AD atrophy subtypes in resting-state functional magnetic resonance imaging intra-network connectivity (INC) and global and nodal network properties. Using a data-driven clustering approach, we identified four AD atrophy subtypes with differences in functional connectivity, accompanied by clinical and biomarker alterations, including a medio-temporal-predominant (S-MT), a limbic-predominant (S-L), a diffuse (S-D), and a mild-atrophy (S-MA) subtype. S-MT and S-D showed INC reduction in the default mode, dorsal attention, visual and limbic network, and a pronounced reduction of “global efficiency” and decrease of the “clustering coefficient” in parietal and temporal lobes. Despite severe atrophy in limbic areas, the S-L exhibited only marginal global network but substantial nodal network failure. S-MA, in contrast, showed limited impairment in clinical and cognitive scores but pronounced global network failure. Our results contribute toward a better understanding of heterogeneity in AD with the detection of distinct differences in functional connectivity networks accompanied by CSF biomarker and cognitive differences in AD subtypes.

CNNs, RNNs, GCNs, and CapsNets have shown significant insights in representation learning and are widely used in various text mining tasks such as large-scale multi-label text classification. Most existing deep models for multi-label text classification consider either the non-consecutive and long-distance semantics or the sequential semantics. However, how to coherently take them into account is still far from studied. In addition, most existing methods treat output labels as independent medoids, ignoring the hierarchical relationships among them, which leads to a substantial loss of useful semantic information. In this paper, we propose a novel hierarchical taxonomy-aware and attentional graph capsule recurrent CNNs framework for large-scale multi-label text classification. Specifically, we first propose to model each document as a word order preserved graph-of-words and normalize it as a corresponding word matrix representation preserving both the non-consecutive, long-distance and local sequential semantics. Then the word matrix is input to the proposed attentional graph capsule recurrent CNNs for effectively learning the semantic features. To leverage the hierarchical relations among the class labels, we propose a hierarchical taxonomy embedding method to learn their representations, and define a novel weighted margin loss by incorporating the label representation similarity. Extensive evaluations on three datasets show that our model significantly improves the performance of large-scale multi-label text classification by comparing with state-of-the-art approaches.

Yuki Motomura,1– 3 Shingo Kitamura,1 Kyoko Nakazaki,1 Kentaro Oba,1,4 Ruri Katsunuma,1,3 Yuri Terasawa,1,5 Akiko Hida,1 Yoshiya Moriguchi,1,3 Kazuo Mishima1,6,7 1Department of Sleep-Wake Disorders, National Institute of Mental Health, National Center of Neurology and Psychiatry, Tokyo, 187-8553, Japan; 2Faculty of Design, Kyushu University, Fukuoka, 815-8540, Japan; 3Integrative Brain Imaging Center, National Center of Neurology and Psychiatry, Tokyo, 187-8553, Japan; 4Department of Human Brain Science, Institute of Development, Aging and Cancer, Tohoku University, Sendai, 980-8575, Japan; 5Department of Psychology, Keio University, Kanagawa, 223-8521, Japan; 6Department of Neuropsychiatry, Akita University Graduate School of Medicine, Akita, 010-8543, Japan; 7International Institute for Integrative Sleep Medicine, University of Tsukuba, Ibaraki, JapanCorrespondence: Kazuo MishimaDepartment of Neuropsychiatry, Akita University Graduate School of Medicine, 1-1-1 Hondo, Akita, 010-8543, JapanTel +81 18 884 6122Fax +81 18 884 6445Email mishima@med.akita-u.ac.jpYuki MotomuraFaculty of Design, Kyushu University, 4-9-1 Shiobaru, Minami-ku, Fukuoka, 815-8540, JapanTel +81 92 533 9493Email motomura@design.kyushu-u.ac.jpPurpose: The thalamus, the region that forms the attentional network and transmits external sensory signals to the entire brain, is important for sleepiness. Herein, we examined the relationship between activity in the thalamus-seed brain network and subjective sleepiness.Materials and Methods: Fifteen healthy male participants underwent an experiment comprising a baseline evaluation and two successive interventions, a 9-day sleep extension followed by 1-night total sleep deprivation. Pre- and post-intervention tests included the Karolinska sleepiness scale and neuroimaging for arterial spin labeling and functional connectivity. We examined the association between subjective sleepiness and the functional magnetic resonance imaging indices.Results: The functional connectivity between the left or right thalamus and various brain regions displayed a significant negative association with subjective sleepiness, and the functional connectivity between the left and right thalamus displayed a significant positive association with subjective sleepiness. The graph theory analysis indicated that the number of positive functional connectivity related to the thalamus showed a strong negative association with subjective sleepiness, and conversely, the number of negative functional connectivity showed a positive association with subjective sleepiness. Arterial spin labeling analysis indicated that the blood flow in both the left and right thalami was significantly negatively associated with subjective sleepiness. Functional connectivity between the anterior cingulate cortex and salience network areas of the left insular cortex, and that between the anterior and posterior cingulate cortices showed a strong positive and negative association with subjective sleepiness, respectively.Conclusion: Subjective sleepiness and the thalamic-cortical network dynamics are strongly related, indicating the application of graph theory to study sleepiness and consciousness. These results also demonstrate that resting functional connectivity largely reflects the “state” of the subject, suggesting that the control of sleep and conscious states is essential when using functional magnetic resonance imaging indices as biomarkers.Keywords: sleep, subjective sleepiness, consciousness, functional connectivity, graph theory, rCBF, ASL

In this article, we investigate problem reduction techniques using stochastic sampling and machine learning to tackle large-scale optimization problems. These techniques heuristically remove decision variables from the problem instance, that are not expected to be part of an optimal solution. First we investigate the use of statistical measures computed from stochastic sampling of feasible solutions compared with features computed directly from the instance data. Two measures are particularly useful for this: 1) a ranking-based measure, favoring decision variables that frequently appear in high-quality solutions; and 2) a correlation-based measure, favoring decision variables that are highly correlated with the objective values. To take this further we develop a machine learning approach, called Machine Learning for Problem Reduction (MLPR), that trains a supervised learning model on easy problem instances for which the optimal solution is known. This gives us a combination of features enabling us to better predict the decision variables that belong to the optimal solution for a given hard problem. We evaluate our approaches using a typical optimization problem on graphs-the maximum weight clique problem. The experimental results show our problem reduction techniques are very effective and can be used to boost the performance of existing solution methods.

This paper addresses the relative position coordinated control problem for spacecraft formation flying under an undirected communication graph, whilst considering mass uncertainties, external disturbances, and limited communication resources. A new event-triggered information transmission mechanism is first presented, where each spacecraft only requires accessing to the states of neighbors intermittently. Subsequently, a novel event-based coordinated control scheme is proposed by combining a smooth adaptive projection rule that confines the parameter estimations to well-defined bounded convex hypercubes. Under the proposed control framework, the information exchange among spacecraft occurs only when the specified event is triggered, thereby significantly reducing the communication load and saving the onboard resources. Furthermore, a positive lower bound on interevent time intervals is guaranteed to exclude Zeno behavior. By virtue of Lyapunov stability analysis and graph theory, it is proved that the relative position tracking errors can converge to small invariant sets around the origin, and that all closed-loop signals are bounded, even in the presence of mass uncertainties and external disturbances. Finally, numerical simulations are given to evaluate the effectiveness and highlight the advantages of the developed control algorithm.

This paper presents a method for allocating active power losses in electric power networks to generators. A technique that uses current distribution factor is used to allocate losses to generator nodes. The core of the allocation scheme is based on graph theory and flows distribution in a network. Losses are only allocated based on the segment of a network that is used for power evacuation. Models of IEEE 14, 39, 57 and 118 test systems in PYPOWER 5.12 were used to test the scheme. It was observed that although the total network losses is minimised when optimal power flow is used for scheduling generation, however that does not translate to minimisation of loss allocation to some generators. The results obtained show that, the scheme can be used to allocate transmission network losses to generation nodes in electric power networks in a fair manner.

Today, Online Social Networks becomes the first choice for businesses to broadcast their campaigns for branding, publicity, strategies, advertising, marketing, social influence and so many other areas. Social Network is a platform for communicating with social actors and Social Media is used by companies for broadcasting information. Online Social Network used by businesses for number of purposes but the primary concern is build new social connections that helps to target most audiences for successful campaign purposes. In OSNs sites the social objects are represented by nodes and the term edge used for connection between nodes under the graph theory. Today Social Network sites have becomes most exploded as compared to traditional sites because of impact of so many influence models over traditional models.Some of popular websites of OSN such as MySpace, Facebook, Flickr, YouTube, Google Video, Orkut, LinkedIn, Live Journal and BlogSpot have great impact on customer when targeting the sales marketing funnel for businesses. Adjacent users sometimes called engaged users tend to have more trust level as compared to random pairs users on the social media sites. Already have so much research that helps to calculate the trust factor using influence modeling. So influence models play a vital role to predict the behavior of the customer that helps to fulfill the goal of the business. The key contribution of this work is study of online social networking models.

Link prediction is challenging, especially based on (scarce) historic data or in cold start scenarios. In this paper, we show how to apply answer set programming (ASP) for formalizing link prediction in feature-rich networks, that is – in particular – using domain knowledge for network (and graph) analysis. We show, that applying ASP for link prediction provides a powerful declarative approach, as exemplified using simple predictors, and demonstrate according explanation generation using ASP. We present the application of the proposed methodological approach for explicative link prediction and analysis with explanation generation using different datasets. Keywords . Explainable AI, social network analysis, link prediction, answer set programming, Prolog, domain knowledge

Recognizing emotion through facial expression has now been widely applied in our daily lives. Therefore, facial expression recognition (FER) is attracting increasing research interests in the field of artificial intelligence and multimedia. With the development of convolutional neural networks (CNN), end-to-end deep learning frameworks for FER have achieved great success on large-scale datasets. However, these works still face the problems of redundant information and data bias, which obviously decrease the performance of FER. In this article, we propose a novel multiscale graph convolutional network (GCN) based on landmark graphs extracted from facial images. The proposed method is evaluated on different popular datasets. The results show that the proposed method outperforms the traditional deep learning frameworks and achieves more stable performance on different datasets.

Receivable financing is the process whereby cash is advanced to firms against receivables their customers have yet to pay: a receivable can be sold to a funder, which immediately gives the firm cash in return for a small percentage of the receivable amount as a fee. Receivable financing has been traditionally handled in a centralized way, where every request is processed by the funder individually and independently of one another. In this work we propose a novel, network-based approach to receivable financing, which enables customers of the same funder to autonomously pay each other as much as possible, and gives benefits to both the funder (reduced cash anticipation and exposure risk) and its customers (smaller fees and lightweight service establishment). Our main contribution consists in providing a principled formulation of the network-based receivable-settlement strategy, and showing how to achieve all algorithmic challenges posed by the design of this strategy. We formulate network-based receivable financing as a novel combinatorial-optimization problem on a multigraph of receivables. We show that the problem is $\mathbf {NP}$ -hard, and devise an exact branch-and-bound algorithm, as well as algorithms to efficiently find effective approximate solutions. Our more efficient algorithms are based on cycle enumeration and selection, and exploit a theoretical characterization in terms of a knapsack problem, as well as a refining strategy that properly adds paths between cycles. We also investigate the real-world issue of avoiding temporary violations of the problem constraints, and design methods for handling it. An extensive experimental evaluation is performed on real receivable data. Results attest the good performance of our methods.

Connectivity analysis between the components of large evolving systems can reveal significant patterns of interaction. The systems can be simulated by topological graph structures. However, such analysis becomes challenging on large and complex graphs. Tasks such as comparing, searching, and summarizing structures, are difficult due to the enormous number of calculations required. For time-varying graphs, the temporal dimension even intensifies the difficulty. In this article, we propose to reduce the complexity of analysis by focusing on subgraphs that are induced by closely related entities. To summarize the diverse structures of subgraphs, we build a supervised layout-based classification model. The main premise is that the graph structures can induce a unique appearance of the layout. In contrast to traditional graph theory-based and contemporary neural network-based methods of graph classification, our approach generates low costs and there is no need to learn informative graph representations. Combined with temporally stable visualizations, we can also facilitate the understanding of sub-structures and the tracking of graph evolution. The method is evaluated on two real-world datasets. The results show that our system is highly effective in carrying out visual-based analytics of large graphs.

Motivation Polypharmacy side effects should be carefully considered for new drug development. However, considering all the complex drug–drug interactions that cause polypharmacy side effects is challenging. Recently, graph neural network (GNN) models have handled these complex interactions successfully and shown great predictive performance. Nevertheless, the GNN models have difficulty providing intelligible factors of the prediction for biomedical and pharmaceutical domain experts. Method A novel approach, graph feature attention network (GFAN), is presented for interpretable prediction of polypharmacy side effects by emphasizing target genes differently. To artificially simulate polypharmacy situations, where two different drugs are taken together, we formulated a node classification problem by using the concept of line graph in graph theory. Results Experiments with benchmark datasets validated interpretability of the GFAN and demonstrated competitive performance with the graph attention network in a previous work. And the specific cases in the polypharmacy side-effect prediction experiments showed that the GFAN model is capable of very sensitively extracting the target genes for each side-effect prediction. Availability and implementation https://github.com/SunjooBang/Polypharmacy-side-effect-prediction.

This article argues that technical leaders in a post-COVID-19 world should adopt stakeholder-oriented management techniques tailored to complex, chaotic, and disordered domains. In part, this is because the unprecedented breadth of the novel coronavirus's impact has required organizations to strengthen their stakeholder orientation, given the second-order, network effects of management decisions during a global pandemic on unseen, unknown constituencies. In part, this is because even our causal models and modes of group sense-making have been affected by the suffering and uncertainty caused by the pandemic. This article draws from ongoing research on complex stakeholder networks, applying recent advances in graph theory to establish that organizations with more resilient, more efficient, more globally connected stakeholder networks better satisfy the claims of their stakeholders, across an array of financial and environmental, social and governance metrics. This research can be extended into postcrisis, chaotic, and disordered domains by adapting the Cynefin framework, from the Welsh word for “habitat,” and applying it to the stakeholder theory literature. This stakeholder network perspective yields surprisingly simple and relevant tools for managers, as shown by a case study of the real-world performance of large US firms during the period of COVID-19’s initial diffusion and impact.

Finding the most suitable co-author is one of the most important ways to conduct effective research in scientific fields. Data science has contributed to achieving this possibility significantly. The present study aims at designing a mathematical model of co-author recommender system in bioinformatics using graph mining techniques and big data applications. The present study employed an applied-developmental research method and a mixed-methods research design. The research population consisted of all scientific products in bioinformatics in the PubMed database. To achieve the research objectives, the most appropriate effective features in choosing a co-author were selected, prioritized, and weighted by experts. Then, they were weighted using graph mining techniques and big data applications. Finally, the mathematical co-author recommender system model in bioinformatics was presented. Data analysis instruments included Expert Choice, Excel, Spark, Scala and Python programming languages in a big data server. The research was conducted in four steps: (1) identifying and prioritizing the criteria effective in choosing a co-author using AHP; (2) determining the correlation degree of articles based on the criteria obtained from step 1 using algorithms and big data applications; (3) developing a mathematical co-author recommender system model; and (4) evaluating the developed mathematical model. Findings showed that the journal titles and citations criteria have the highest weight while the abstract has the lowest weight in the mathematical co-author recommender system model. The accuracy of the proposed model was 72.26. It was concluded that using content-based features and expert opinions have high potentials in recommending the most appropriate co-authors. It is expected that the proposed co-author recommender system model can provide appropriate recommendations for choosing co-authors on various fields in different contexts of scientific information. The most important innovation of this model is the use of a combination of expert opinions and systemic weights, which can accelerate the finding of co-authors and consequently saving time and achieving a greater quality of scientific products.

Xiang-Xin Xing,1,2,* Xu-Yun Hua,3,4,* Mou-Xiong Zheng,3,* Jia-Jia Wu,2 Bei-Bei Huo,1,2 Jie Ma,1,2 Zhen-Zhen Ma,1,2 Si-Si Li,1,2 Jian-Guang Xu1 1School of Rehabilitation Science, Shanghai University of Traditional Chinese Medicine, Shanghai, People’s Republic of China; 2Department of Rehabilitation Medicine, Yueyang Hospital, Shanghai University of Traditional Chinese Medicine, Shanghai, People’s Republic of China; 3Department of Traumatology and Orthopedics, Yueyang Hospital, Shanghai University of Traditional Chinese Medicine, Shanghai, People’s Republic of China; 4Yangzhi Rehabilitation Hospital, Tongji University, Shanghai, People’s Republic of China*These authors contributed equally to this workCorrespondence: Jian-Guang XuSchool of Rehabilitation Science, Shanghai University of Traditional Chinese Medicine, No. 1200 Cailun Road, Shanghai, People’s Republic of ChinaTel +86-21-51322091Fax +86-21-51322042Email xjg@shutcm.edu.cnIntroduction: Numerous resting-state functional magnetic resonance imaging (fMRI) researches have indicated that large-scale functional and structural remodeling occurs in the whole brain despite an intact sensorimotor network after carpal tunnel syndrome (CTS). Investigators aimed to explore alterations of the global and nodal properties that occur in the whole brain network of patients with CTS based on topographic theory.Methods: Standard-compliant fMRI data were collected from 27 patients with CTS in bilateral hands and 19 healthy control subjects in this cross-sectional study. The statistics based on brain networks were calculated the differences between the patients and the healthy. Several topological properties were computed, such as the small-worldness, nodal clustering coefficient, characteristic path length, and degree centrality.Results: Compared to those of the healthy controls, the global properties of the CTS group exhibited a decreased characteristic path length. Changes in the local-level properties included a decreased nodal clustering coefficient in 6 separate brain regions and significantly different degree centrality in several brain regions that were related to sensorimotor function and pain.Discussion: The study suggested that CTS reinforces global connections and makes their networks more random. The changed nodal properties were affiliated with basal ganglia-thalamo-cortical circuits and the pain matrix. These results provided new insights for improving our understanding of abnormal topological theory in relation to the functional brain networks of CTS patients.Perspective: This article presents that the CTS patients’ brain with a higher global efficiency. And the significant alterations in several brain regions which are more related to pain and motor processes. The results provided effective complements to the neural mechanisms underlying CTS.Keywords: small-worldness, nodal clustering coefficient, degree centrality, pain, circuits

One of the important approaches of handling data heterogeneity in multimodal data clustering is modeling each modality using a separate similarity graph. Information from the multiple graphs is integrated by combining them into a unified graph. A major challenge here is how to preserve cluster information while removing noise from individual graphs. In this regard, a novel algorithm, termed as CoALa, is proposed that integrates noise-free approximations of multiple similarity graphs. The proposed method first approximates a graph using the most informative eigenpairs of its Laplacian which contain cluster information. The approximate Laplacians are then integrated for the construction of a low-rank subspace that best preserves overall cluster information of multiple graphs. However, this approximate subspace differs from the full-rank subspace which integrates information from all the eigenpairs of each Laplacian. Matrix perturbation theory is used to theoretically evaluate how far approximate subspace deviates from the full-rank one for a given value of approximation rank. Finally, spectral clustering is performed on the approximate subspace to identify the clusters. Experimental results on several real-life cancer and benchmark data sets demonstrate that the proposed algorithm significantly and consistently outperforms state-of-the-art integrative clustering approaches.

Application scientists often employ feature tracking algorithms to capture the temporal evolution of various features in their simulation data. However, as the complexity of the scientific features is increasing with the advanced simulation modeling techniques, quantification of reliability of the feature tracking algorithms is becoming important. One of the desired requirements for any robust feature tracking algorithm is to estimate its confidence during each tracking step so that the results obtained can be interpreted without any ambiguity. To address this, we develop a confidence-guided feature tracking algorithm that allows reliable tracking of user-selected features and presents the tracking dynamics using a graph-based visualization along with the spatial visualization of the tracked feature. The efficacy of the proposed method is demonstrated by applying it to two scientific datasets containing different types of time-varying features.

The mapping of the wiring diagrams of neural circuits promises to allow us to link structure and function of neural networks. Current approaches to analyzing connectomes rely mainly on graph-theoretical tools, but these may downplay the complex nonlinear dynamics of single neurons and networks, and the way networks respond to their inputs. Here, we measure the functional similarity of simulated networks of neurons, by quantifying the similitude of their spiking patterns in response to the same stimuli. We find that common graph theory metrics convey little information about the similarity of networks’ responses. Instead, we learn a functional metric between networks based on their synaptic differences, and show that it accurately predicts the similarity of novel networks, for a wide range of stimuli. We then show that a sparse set of architectural features - the sum of synaptic inputs that each neuron receives and the sum of each neuron’s synaptic outputs - predicts the functional similarity of networks of up to 100 cells, with high accuracy. We thus suggest new architectural design principles that shape the function of neural networks, which conform with experimental evidence of homeostatic mechanisms.