Network Representation Method Research Articles

Notes on Bus User Assignment Problem Using Section Network Representation Method

Notes on Bus User Assignment Problem Using Section Network Representation Method

Ordinal Pattern Tree: A New Representation Method for Brain Network Analysis.

Brain networks, describing the functional or structural interactions of brain with graph theory, have been widely used for brain imaging analysis. Currently, several network representation methods have been developed for describing and analyzing brain networks. However, most of these methods ignored the valuable weighted information of the edges in brain networks. In this paper, we propose a new representation method (i.e., ordinal pattern tree) for brain network analysis. Compared with the existing network representation methods, the proposed ordinal pattern tree (OPT) can not only leverage the weighted information of the edges but also express the hierarchical relationships of nodes in brain networks. On OPT, nodes are connected by ordinal edges which are constructed by using the ordinal pattern relationships of weighted edges. We represent brain networks as OPTs and further develop a new graph kernel called optimal transport (OT) based ordinal pattern tree (OT-OPT) kernel to measure the similarity between paired brain networks. In OT-OPT kernel, the OT distances are used to calculate the transport costs between the nodes on the OPTs. Based on these OT distances, we use exponential function to calculate OT-OPT kernel which is proved to be positive definite. To evaluate the effectiveness of the proposed method, we perform classification and regression experiments on ADHD-200, ABIDE and ADNI datasets. The experimental results demonstrate that our proposed method outperforms the state-of-the-art graph methods in the classification and regression tasks.

Trajectory Data-Driven Network Representation for Traffic State Prediction using Deep Learning

In this study, we propose a trajectory data-driven network representation method, specifically leveraging directional statistics. This approach allows us to extract major intersections and define links from observed trajectories, thereby mitigating the reliance on existing network data and map matching. We apply Graph Convolutional Networks and Long-Short Term Memory models to the trajectory data-driven network representation, suggesting the potential for fast and accurate traffic state prediction. The results imply significant reduction in computational complexity while demonstrating promising prediction accuracy. Our proposed method offers a valuable approach for analyzing and modeling transportation networks using real-world trajectory data, providing insights into traffic patterns and facilitating the exploration of more efficient traffic management strategies.

Variational Graph Autoencoder with Adversarial Mutual Information Learning for Network Representation Learning

With the success of Graph Neural Network (GNN) in network data, some GNN-based representation learning methods for networks have emerged recently. Variational Graph Autoencoder (VGAE) is a basic GNN framework for network representation. Its purpose is to well preserve the topology and node attribute information of the network to learn node representation, but it only reconstructs network topology, and does not consider the reconstruction of node features. This strategy will make node representation can not well reserve node features information, impairing the ability of the VGAE method to learn higher quality representations. To solve this problem, we arise a new network representation method to improve the VGAE method for well retaining both node features and network structure information. The method utilizes adversarial mutual information learning to maximize the mutual information (MI) of node features and node representations during the encoding process of the variational autoencoder, which forces the variational encoder to get the representation containing the most informative node features. The method consists of three parts: a variational graph autoencoder includes a variational encoder (MI generator (G)) and a decoder, a positive MI sample module (maximizing MI module), and an MI discriminator (D). Furthermore, we explain why maximizing MI between node features and node representation can reconstruct node attributes. Finally, we conduct experiments on seven public representative datasets for nodes classification, nodes clustering, and graph visualization tasks. Experimental results demonstrate that the proposed algorithm significantly outperforms current popular network representation algorithms on these tasks. The best improvement is 17.13% than the VGAE method.

Topological Relationship‐Based Flow Direction Modeling: Mesh‐Independent River Networks Representation

AbstractRiver networks are important features in surface hydrology. However, accurately representing river networks in spatially distributed hydrologic and Earth system models is often sensitive to the model's spatial resolution. Specifically, river networks are often misrepresented because of the mismatch between the model's spatial resolution and river network details, resulting in significant uncertainty in the projected flow direction. In this study, we developed a topological relationship‐based river network representation method for spatially distributed hydrologic models. This novel method uses (a) graph theory algorithms to simplify real‐world vector‐based river networks and assist in mesh generation; and (b) a topological relationship‐based method to reconstruct conceptual river networks. The main advantages of our method are that (a) it combines the strengths of vector‐based and DEM raster‐based river network extraction methods; and (b) it is mesh‐independent and can be applied to both structured and unstructured meshes. This method paves a path for advanced terrain analysis and hydrologic modeling across different scales.

Positive-Unlabeled Learning for Network Link Prediction

Link prediction is an important problem in network data mining, which is dedicated to predicting the potential relationship between nodes in the network. Normally, network link prediction based on supervised classification will be trained on a dataset consisting of a set of positive samples and a set of negative samples. However, well-labeled training datasets with positive and negative annotations are always inadequate in real-world scenarios, and the datasets contain a large number of unlabeled samples that may hinder the performance of the model. To address this problem, we propose a positive-unlabeled learning framework with network representation for network link prediction only using positive samples and unlabeled samples. We first learn representation vectors of nodes using a network representation method. Next, we concatenate representation vectors of node pairs and then feed them into different classifiers to predict whether the link exists or not. To alleviate data imbalance and enhance the prediction precision, we adopt three types of positive-unlabeled (PU) learning strategies to improve the prediction performance using traditional classifier estimation, bagging strategy and reliable negative sampling. We conduct experiments on three datasets to compare different PU learning methods and discuss their influence on the prediction results. The experimental results demonstrate that PU learning has a positive impact on predictive performances and the promotion effects vary with different network structures.

Finding Asymptomatic Spreaders in a COVID-19 Transmission Network by Graph Attention Networks.

In the COVID-19 epidemic the mildly symptomatic and asymptomatic infections generate a substantial portion of virus spread; these undetected individuals make it difficult to assess the effectiveness of preventive measures as most epidemic prevention strategies are based on the detected data. Effectively identifying the undetected infections in local transmission will be of great help in COVID-19 control. In this work, we propose an RNA virus transmission network representation model based on graph attention networks (RVTR); this model is constructed using the principle of natural language processing to learn the information of gene sequence and using a graph attention network to catch the topological character of COVID-19 transmission networks. Since SARS-CoV-2 will mutate when it spreads, our approach makes use of graph context loss function, which can reflect that the genetic sequence of infections with close spreading relation will be more similar than those with a long distance, to train our model. Our approach shows its ability to find asymptomatic spreaders both on simulated and real COVID-19 datasets and performs better when compared with other network representation and feature extraction methods.

Network representation learning method embedding linear and nonlinear network structures

With the rapid development of neural networks, much attention has been focused on network embedding for complex network data, which aims to learn low-dimensional embedding of nodes in the network and how to effectively apply learned network representations to various graph-based analytical tasks. Two typical models exist namely the shallow random walk network representation method and deep learning models such as graph convolution networks (GCNs). The former one can be used to capture the linear structure of the network using depth-first search (DFS) and width-first search (BFS), whereas Hierarchical GCN (HGCN) is an unsupervised graph embedding that can be used to describe the global nonlinear structure of the network via aggregating node information. However, the two existing kinds of models cannot simultaneously capture the nonlinear and linear structure information of nodes. Thus, the nodal characteristics of nonlinear and linear structures are explored in this paper, and an unsupervised representation method based on HGCN that joins learning of shallow and deep models is proposed. Experiments on node classification and dimension reduction visualization are carried out on citation, language, and traffic networks. The results show that, compared with the existing shallow network representation model and deep network model, the proposed model achieves better performances in terms of micro-F1, macro-F1 and accuracy scores.

Prediction of the disease causal genes based on heterogeneous network and multi-feature combination method

At present, the prediction of disease causal genes is mainly based on heterogeneous. Research shows that heterogeneous network contains more information and have better prediction results. In this paper, we constructed a heterogeneous network including four node types of disease, gene, phenotype and gene ontology. On this basis, we use a machine learning algorithm to predict disease-causing genes. The algorithm is divided into three steps: preprocess and training sample extraction, features extraction and combination, model training and prediction. In the process of feature extraction and combination, by using network representation method, the representation vectors of nodes are generated as the embedding features of the nodes. We also extracted the structural features of each node in the network and then the embedding features and structure features are combined. The results of training and prediction show that the prediction algorithm based on all features combined together achieves the best prediction performance. Moreover, the combination of each network representation method’s embedding features and structural features has also achieved performance improvement. In the process of training samples extraction, we propose three improvement directions according to the network structure and data set distribution. Firstly, a positive sample algorithm based on network connectivity is proposed, we try to keep the connectivity of the whole heterogeneous graph in the sampling process to avoid the negative impact of embedding features’ extraction. Moreover, the influence of sample sampling ratio on experimental results was tested in the range of 0–1 with step size of 0.1. The influence of different proportion of positive and negative samples on the results was also tested. These improvements are intended to enhance the balance and robustness of the method. When the positive sample ratio is 0.1 and the proportion of negative and positive samples is 3, the model achieves the optimal result, and its AUC value and accuracy are 0.9887% and 94.55%, respectively, which are significantly higher than other models.

HEPre: Click frequency prediction of applications based on heterogeneous information network embedding

Owing the continuous enrichment of mobile application resources, mobile applications carry almost all user behaviors and preferences. The analysis of user behavior regarding mobile terminals has become an important research direction. The frequency with which users click on mobile applications reflects their preferences to a certain extent. In this study, we propose a mobile application click-frequency prediction model based on heterogeneous information network representation. This model first constructs a heterogeneous information network between users’ mobile devices and mobile applications. To generate a meaningful sequence of network-embedded nodes, we perform a random walk on a specified meta-path. Finally, the prediction of users’ mobile application click frequency is completed using representation fusion and matrix factorization. Experiments show that our method outperforms other baseline methods in terms of the mean absolute error and root mean square error. Therefore, the application of a heterogeneous information network representation method to the prediction model is effective. This study is significant to the behavior research of mobile terminal users.

LDNM: A General Web Service Classification Framework via Deep Fusion of Structured and Unstructured Features

Classifying Web services plays a critical role in several fundamental service management tasks, such as service discovery, selection, ranking, and recommendation. However, traditional Web service classification approaches usually difficult to dispose unstructured sparse documents and underutilize the rich network relations. The consideration of multiple document representation schemes can ameliorate the former problem, whereas an appropriate network representation method could be a positive solution to the latter problem. In this paper, we propose a general Web service classification framework via deep fusion of structured and unstructured features, named LDNM. Firstly, we transform each service document into feature vectors by using two document representation methods: topic distribution based on LDA, and neural-network-based document embedding model known as Doc2vec. Then we obtain structured representation vectors which stem from service invoking and tagging graphs by applying Node2vec. Finally, we fuse these features and train a service classifier by using an MLP neural network. Comprehensive experiments are conducted on real-world datasets to demonstrate the effectiveness of the proposed approach.

Modelling of integrated scheduling problem of capacitated equipment systems with a multi-lane road network.

The coordination of different container-handling equipment is an important method for improving the overall efficiency of automated container terminals. In the real terminal, we should consider many real-life issues, such as the equipment capacity, the equipment collision, changing lanes in the multi-lane road, and choosing one of container-handling lanes for each container. This paper proposes the integrated scheduling problem of three container-handling equipment with the capacity constraint and the dual-cycle strategy, for simultaneously solving the equipment scheduling problem, the assignment problem of the container-handling lane and the conflict-free route planning problem of automated guided vehicles (AGVs). With the objective of minimizing the ship’s berth time, we propose a mixed-integer programming model based on the space-time network representation method and two bilevel optimization algorithms based on conflict resolution rules. Finally, numerical experiments are conducted to verify the effectiveness of the proposed model and two bilevel optimization algorithms.

Disentangling age- and disease-related alterations in schizophrenia brain network using structural equation modeling: A graph theoretical study based on minimum spanning tree.

Functional brain networks have been shown to undergo fundamental changes associated with aging or schizophrenia. However, the mechanism of how these factors exert influences jointly or interactively on brain networks remains elusive. A unified recognition of connectomic alteration patterns was also hampered by heterogeneities in network construction and thresholding methods. Recently, an unbiased network representation method regardless of network thresholding, so called minimal spanning tree algorithm, has been applied to study the critical skeleton of the brain network. In this study, we aimed to use minimum spanning tree (MST) as an unbiased network reconstruction and employed structural equation modeling (SEM) to unravel intertwined relationships among multiple phenotypic and connectomic variables in schizophrenia. First, we examined global and local brain network properties in 40 healthy subjects and 40 schizophrenic patients aged 21–55 using resting‐state functional magnetic resonance imaging (rs‐fMRI). Global network alterations are measured by graph theoretical metrics of MSTs and a connectivity‐transitivity two‐dimensional approach was proposed to characterize nodal roles. We found that networks of schizophrenic patients exhibited a more star‐like global structure compared to controls, indicating excessive integration, and a loss of regional transitivity in the dorsal frontal cortex (corrected p <.05). Regional analysis of MST network topology revealed that schizophrenia patients had more network hubs in frontal regions, which may be linked to the “overloading” hypothesis. Furthermore, using SEM, we found that the level of MST integration mediated the influence of age on negative symptom severity (indirect effect 95% CI [0.026, 0.449]). These findings highlighted an altered network skeleton in schizophrenia and suggested that aging‐related enhancement of network integration may undermine functional specialization of distinct neural systems and result in aggravated schizophrenic symptoms.

A Heterogeneous Network Representation Method Based on Variational Inference and Meta-Path Decomposition

A Heterogeneous Network Representation Method Based on Variational Inference and Meta-Path Decomposition

A network representation learning method based on topology

In the network, nodes with similar neighborhood topology often have similar functions and play similar roles. Accurately identifying the role of nodes is the key to our understanding of the network and then facilitates the completion of subsequent tasks. This paper proposes a network representation method based on the Neighborhood Topology Feature (NTF), which learns the latent representations of nodes based on their neighborhood topologies. NTF adopts the idea of energy dissipation and introduces the concept of energy level to measure the influence of neighboring nodes on the central node, and on this basis constructs the influence subgraph of the central node. NTF comprehensively uses the absolute and relative features of the neighboring nodes to describe the neighborhood topology of the central node hierarchically to achieve a better learning effect. It can effectively reduce the interference of noise nodes by measuring the node pairs similarity on the influence subgraph. The experimental results show that NTF performs better than the existing methods on the public datasets, and it also achieves excellent results on the real data of the actual project.

Wi-Fi-Based Location-Independent Human Activity Recognition via Meta Learning

Wi-Fi-based device-free human activity recognition has recently become a vital underpinning for various emerging applications, ranging from the Internet of Things (IoT) to Human–Computer Interaction (HCI). Although this technology has been successfully demonstrated for location-dependent sensing, it relies on sufficient data samples for large-scale sensing, which is enormously labor-intensive and time-consuming. However, in real-world applications, location-independent sensing is crucial and indispensable. Therefore, how to alleviate adverse effects on recognition accuracy caused by location variations with the limited dataset is still an open question. To address this concern, we present a location-independent human activity recognition system based on Wi-Fi named WiLiMetaSensing. Specifically, we first leverage a Convolutional Neural Network and Long Short-Term Memory (CNN-LSTM) feature representation method to focus on location-independent characteristics. Then, in order to well transfer the model across different positions with limited data samples, a metric learning-based activity recognition method is proposed. Consequently, not only the generalization ability but also the transferable capability of the model would be significantly promoted. To fully validate the feasibility of the presented approach, extensive experiments have been conducted in an office with 24 testing locations. The evaluation results demonstrate that our method can achieve more than 90% in location-independent human activity recognition accuracy. More importantly, it can adapt well to the data samples with a small number of subcarriers and a low sampling rate.

A content-sensitive citation representation approach for citation recommendation

Citation recommendation systems mainly help researchers find the lists of references that related to their interests effectively and automatically. The existing approaches face the issues of data sparsity and high-dimensional in large-scale bibliographic network representation, which hinder the citation recommendation performance. To address these problems, we proposed a Content-Sensitive citation representation approach for Citation Recommendation, named CSCR. Firstly, the Doc2vec model is used to generate a paper embedding according to paper content. Then, utilizing the similarity between the paper content embeddings to select the assumed neighbours of the target paper, append the auxiliary links between target paper and its new neighbours in the bibliographic network. Thirdly, distributed network representation method is implemented on appended bibliographic network to obtain the paper node embedding, which can learn interpretable lower dimension embedding for paper nodes. Finally, the embedding vectors of these papers can be used to conduct citation recommendation. Experimental results show that the proposed approach significantly outperforms other benchmark methods in Normalized Discounted Cumulative Gain (NDCG) and the positive rate (Recall).

Learning Universal Network Representation via Link Prediction by Graph Convolutional Neural Network

Network representation learning algorithms, which aim at automatically encoding graphs into low-dimensional vector representations with a variety of node similarity definitions, have a wide range of downstream applications. Most existing methods either have low accuracies in downstream tasks or a very limited application field, such as article classification in citation networks. In this paper, we propose a novel network representation method, named Link Prediction based Network Representation (LPNR), which generalizes the latest graph neural network and optimizes a carefully designed objective function that preserves linkage structures. LPNR can not only learn meaningful node representations that achieve competitive accuracy in node centrality measurement and community detection but also achieve high accuracy in the link prediction task. Experiments prove the effectiveness of LPNR on three real-world networks. With the mini-batch and fixed sampling strategy, LPNR can learn the embedding of large graphs in a few hours.

Dynamic community discovery via common subspace projection

Detecting communities of highly internal and low external interactions in dynamically evolving networks has become increasingly important owing to its wide applications in divers fields. Conventional solutions based on static community detection approaches treat each snapshot of dynamic networks independently, which may fragment communities in time (Aynaud T and Guillaume J L 2010 8th Int. Symp. on Modeling and Optimization in Mobile, Ad Hoc, and Wireless Networks (IEEE) pp 513–9), resulting in the problem of instability. In this work, we develop a novel dynamic community detection algorithm by leveraging the encoding–decoding scheme present in a succinct network representation method to reconstruct each snapshot via a common low-dimensional subspace, which can remove non-significant links and highlight the community structures, resulting in the mitigation of community instability to a large degree. We conduct experiments on simulated data and real social networking data with ground truths (GT) and compare the proposed method with several baselines. Our method is shown to be more stable without missing communities and more effective than the baselines with competitive performance. The distribution of community size in our method is more in line with the real distribution than those of the baselines at the same time.

Identifying Disease Related Genes by Network Representation and Convolutional Neural Network.

The identification of disease related genes plays essential roles in bioinformatics. To achieve this, many powerful machine learning methods have been proposed from various computational aspects, such as biological network analysis, classification, regression, deep learning, etc. Among them, deep learning based methods have gained big success in identifying disease related genes in terms of higher accuracy and efficiency. However, these methods rarely handle the following two issues very well, which are (1) the multifunctions of many genes; and (2) the scale-free property of biological networks. To overcome these, we propose a novel network representation method to transfer individual vertices together with their surrounding topological structures into image-like datasets. It takes each node-induced sub-network as a represented candidate, and adds its environmental characteristics to generate a low-dimensional space as its representation. This image-like datasets can be applied directly in a Convolutional Neural Network-based method for identifying cancer-related genes. The numerical experiments show that the proposed method can achieve the AUC value at 0.9256 in a single network and at 0.9452 in multiple networks, which outperforms many existing methods.