Search

Over the past few decades, the study of leadership theory has expanded across various disciplines, delving into the intricacies of human behavior and defining the roles of individuals within organizations. Its primary objective is to identify leaders who play significant roles in the communication flow. In addition, behavioral theory posits that leaders can be distinguished based on their daily conduct, while social network analysis provides valuable insights into behavioral patterns. Our study investigates five and six types of social networks frequently observed in different organizations. This study is conducted using datasets we collected from an IT company and public datasets collected from a manufacturing company for the thorough evaluation of prediction performance. We leverage PageRank and effective word embedding techniques to obtain novel features. State-of-the-art performance is obtained using various statistical machine learning methods, graph convolutional networks (GCN), automated machine learning (AutoML), and explainable artificial intelligence (XAI). More specifically, our approach can achieve state-of-the-art performance with an accuracy close to 90 % for leaders identification with data from projects of different types. This investigation contributes to the establishment of sustainable leadership practices by aiding organizations in retaining their leadership talent. [ABSTRACT FROM AUTHOR]

PageRank can be viewed as a hyperlink-based method for estimating the importance of nodes in a network, and has attracted a lot of attention from researchers. In this paper, we first propose an extrapolation procedure for PageRank vector estimation based on the linear combination of the Ritz values computed from the Hessenberg process and eigenvectors from the classical power method. In order to improve the convergence rate of the PageRank computations, the extrapolation procedure is introduced into the Hessenberg-type algorithm by Gu et al. (Numer. Algorithms, 89(4): 1845–1863, 2022), then a new algorithm is derived and named Hessenberg-extrapolation algorithm whose convergence and construction are investigated in detail. In addition, we further analyze the convergence of the Hessenberg-type algorithm and discuss the relationship between the approximate vectors of the Hessenberg-type algorithm and the Arnoldi-type algorithm. Numerical experiments on several examples demonstrate that our presented algorithm has great potential in PageRank approximation. [ABSTRACT FROM AUTHOR]

Precisely recommending relevant items to users is a challenging task because the user’s rating can be influenced by various features. Therefore, social recommender systems have recently been introduced to leverage both the user-item interaction graph and the user-user social relation graph for more accurate rating predictions. Moreover, as graph neural networks (GNN) have demonstrated superior performance in graph representation learning, several algorithms have been developed to incorporate GNN into social recommender systems. However, when the sizes of the social graph and user-item graph are very large, the computational demands of existing GNN-based social recommender systems for aggregating user and item nodes becomes the primary bottleneck. In this paper, we develop a novel lightweight GNN-based social recommender system (called LiteGSR) that effectively reduces the computational overhead associated with aggregation operations while maintaining accuracy. To achieve this, we propose a new approach for refining the social graph by utilizing PageRank-based centrality scores of users and adapting representative virtual users in the user-item graph. Experimental results demonstrate that our new social recommender systems outperform existing state-of-the-art recommender systems in both accuracy and training time. [ABSTRACT FROM AUTHOR]

In recent years, the PageRank algorithm has garnered significant attention due to its crucial role in search engine technologies and its applications across various scientific fields. It is well-known that the power method is a classical method for computing PageRank. However, there is a pressing demand for alternative approaches that can address its limitations and enhance its efficiency. Specifically, the power method converges very slowly when the damping factor is close to 1. To address this challenge, this paper introduces a modified multi-step splitting iteration approach for accelerating PageRank computations. Furthermore, we present two variants for computing PageRank, which are variants of the modified multi-step splitting iteration approach, specifically utilizing the thick restarted Arnoldi and adaptively accelerated Arnoldi methods. We provide detailed discussions on the construction and theoretical convergence results of these two approaches. Extensive experiments using large test matrices demonstrate the significant performance improvements achieved by our proposed algorithms. [ABSTRACT FROM AUTHOR]

Abstract With the ongoing biodiversity crisis, identifying which species are of particular importance to prevent the extinction of other species has become a pressing issue. However, most approaches to detect these important species are made at a local (i.e, community) level, without considering the potential effect of species dispersion in a landscape. As habitat fragmentation has important effects on biodiversity, we need methods to better assess how local and spatial processes interact to determine important species and therefore better inform conservation efforts. We present a modified PageRank algorithm to determine the importance of species in meta‐communities. Species importance is defined as the ability of species in the meta‐community to spread nutrients within two sets of networks: food webs that depict local trophic interactions and landscape networks representing the movement of species across different habitat patches. We show that dispersal and trophic links jointly determine the importance of the different species, both at the local scale (within habitat patches) and at larger scales (landscape). More precisely, we observed that (i) what is considered an important species changes between isolated communities and meta‐communities and (ii) the importance of a species in a meta‐community depends on the position of its habitat patch in the landscape network. The importance of a species is influenced by both intrinsic factors (dispersal capacity, trophic position) and extrinsic factors (position of the patch in the landscape network). Our results stress the need for a larger‐scale consideration of space in the identification of important species.

Social media technology affects the learning paradigm change towards social media-based learning, known as social e-learning. Social e-learning regards a person as a center of learning, dubbed people-centered learning. Here, people are encouraged to interact or communicate with others and produce their learning content. This work attempted to provide a solution model for rural e-learning social learning empowered with intelligent web technology. The proposed social e-learning includes several modules for development, such as personal space, collaboration space, and communication space modules. It also leverages intelligent web technologies currently implemented in today’s social media applications, such as article search, article recommendations, friendship recommendations, and document classification. In the searching module, the PageRank method was used to calculate the relevance score to determine the rating of the documents or articles. The similarity-based element calculation method was utilized to create articles’ suggestions and recommendations. The naïve Bayes algorithm, decision tree, and neural network were compared to find the best solution for article classification in agriculture, fisheries, animal husbandry, and plantations. When comparing these three algorithms, the result showed that the neural network was the most accurate classification, reaching 95.2% accuracy. A clustering algorithm, namely robust clustering using links (ROCK), was utilized for rural friendship recommendation. Thus, these algorithms (the PageRank, the similarity-based element, neural network, and the ROCK) were suitable and recommended for supporting intelligent web paradigms in social e-learning applications.

Google employs PageRank to rank web pages, determining the order in which search results are presented to users based on their queries. PageRank is primarily utilized for directed networks, although there are instances where it is also applied to undirected networks. In this paper, we have applied PageRank to undirected networks, showing that a vertex's PageRank relies on its initial value, often referred to as an intrinsic, non-network contribution. We have analytically proved that when the initial value of vertices is either proportional to their degrees or set to zero, the PageRank values of the vertices become directly proportional to their degrees. Simulated and empirical data are employed to bolster our research findings. Additionally, we have investigated the impact of initial values on PageRank localization. [ABSTRACT FROM AUTHOR]

Many ranking algorithms and metrics have been proposed to identify high-impact papers. Both the direct citation counts and the network-based PageRank-like algorithms are commonly used. Ideally, the more complete the data on the citation network, the more informative the ranking. However, obtaining more data on citation relations is often costly and challenging. In some cases, obtaining the citation counts can be relatively simple. In this paper, we look into using the additional citation counts but without additional citation relations to form more informative metrics for identifying high-impact papers. As an example, we propose enhancing the original PageRank algorithm by combining the local citation network with the additional citation counts from a more complete data source. We apply this enhanced method to American Physical Society (APS) papers to verify its effectiveness. The results indicate that the proposed ranking algorithm is robust against missing data and can improve the identification of high-quality papers. This shows that it is possible to enhance the effectiveness of a network-based metric calculated on a relatively small citation network by including only the additional data of the citation counts, without the additional citation relations. [ABSTRACT FROM AUTHOR]

With the ongoing biodiversity crisis, identifying which species are of particular importance to prevent the extinction of other species has become a pressing issue. However, most approaches to detect these important species are made at a local (i.e, community) level, without considering the potential effect of species dispersion in a landscape. As habitat fragmentation has important effects on biodiversity, we need methods to better assess how local and spatial processes interact to determine important species and therefore better inform conservation efforts.We present a modified PageRank algorithm to determine the importance of species in meta‐communities. Species importance is defined as the ability of species in the meta‐community to spread nutrients within two sets of networks: food webs that depict local trophic interactions and landscape networks representing the movement of species across different habitat patches.We show that dispersal and trophic links jointly determine the importance of the different species, both at the local scale (within habitat patches) and at larger scales (landscape). More precisely, we observed that (i) what is considered an important species changes between isolated communities and meta‐communities and (ii) the importance of a species in a meta‐community depends on the position of its habitat patch in the landscape network.The importance of a species is influenced by both intrinsic factors (dispersal capacity, trophic position) and extrinsic factors (position of the patch in the landscape network). Our results stress the need for a larger‐scale consideration of space in the identification of important species. [ABSTRACT FROM AUTHOR]

A network evolution with predicted tail and extremal indices of PageRank and the Max-Linear Model used as node influence indices in random graphs is considered. The tail index shows a heaviness of the distribution tail. The extremal index is a measure of clustering (or local dependence) of the stochastic process. The cluster implies a set of consecutive exceedances of the process over a sufficiently high threshold. Our recent results concerning sums and maxima of non-stationary random length sequences of regularly varying random variables are extended to random graphs. Starting with a set of connected stationary seed communities as a hot spot and ranking them with regard to their tail indices, the tail and extremal indices of new nodes that are appended to the network may be determined. This procedure allows us to predict a temporal network evolution in terms of tail and extremal indices. The extremal index determines limiting distributions of a maximum of the PageRank and the Max-Linear Model of newly attached nodes. The exposition is provided by algorithms and examples. To validate our theoretical results, our simulation and real data study concerning a linear preferential attachment as a tool for network growth are provided. [ABSTRACT FROM AUTHOR]

The growth in academic and scientific publications has increased very rapidly. Researchers must choose a representative and significant literature for their research, which has become challenging worldwide. Usually, the paper citation number indicates this paper's potential influence and importance. However, this standard metric of citation numbers is not suitable to assess the popularity and significance of recently published papers. To address this challenge, this study presents an effective prediction method called FutureCite to predict the future citation level of research articles. FutureCite integrates machine learning with text and graph mining techniques, leveraging their abilities in classification, datasets in-depth analysis, and feature extraction. FutureCite aims to predict future citation levels of research articles applying a multilabel classification approach. FutureCite can extract significant semantic features and capture the interconnection relationships found in scientific articles during feature extraction using textual content, citation networks, and metadata as feature resources. This study's objective is to contribute to the advancement of effective approaches impacting the citation counts in scientific publications by enhancing the precision of future citations. We conducted several experiments using a comprehensive publication dataset to evaluate our method and determine the impact of using a variety of machine learning algorithms. FutureCite demonstrated its robustness and efficiency and showed promising results based on different evaluation metrics. Using the FutureCite model has significant implications for improving the researchers' ability to determine targeted literature for their research and better understand the potential impact of research publications. [ABSTRACT FROM AUTHOR]

In this study, we explore how transformer models, which are known for their attention mechanisms, can improve pathogen prediction in pastured poultry farming. By combining farm management practices with microbiome data, our model outperforms traditional prediction methods in terms of the F1 score—an evaluation metric for model performance—thus fulfilling an essential need in predictive microbiology. Additionally, the emphasis is on making our model's predictions explainable. We introduce a novel approach for identifying feature importance using the model's attention matrix and the PageRank algorithm, offering insights that enhance our comprehension of established techniques such as DeepLIFT. Our results showcase the efficacy of transformer models in pathogen prediction for food safety and mark a noteworthy contribution to the progress of explainable AI within the biomedical sciences. This study sheds light on the impact of effective farm management practices and highlights the importance of technological advancements in ensuring food safety. [ABSTRACT FROM AUTHOR]

This research leverages Memgraph, an open-source graph database, to analyze graph-based network data and apply Graph Neural Networks (GNNs) for a detailed classification of cyberattack tactics categorized by the MITRE ATT&CK framework. As part of graph characterization, the page rank, degree centrality, betweenness centrality, and Katz centrality are presented. Node classification is utilized to categorize network entities based on their role in the traffic. Graph-theoretic features such as in-degree, out-degree, PageRank, and Katz centrality were used in node classification to ensure that the model captures the structure of the graph. The study utilizes the UWF-ZeekDataFall22 dataset, a newly created dataset which consists of labeled network logs from the University of West Florida's Cyber Range. The uniqueness of this study is that it uses the power of combining graph-based characterization or analysis with machine learning to enhance the understanding and visualization of cyber threats, thereby improving the network security measures. [ABSTRACT FROM AUTHOR]

The Classifier Chains (CC) method is an effective method for multi-label classification, with its performance significantly contingent on the label order. However, most existing algorithms focus on determining the label order by considering label correlations as symmetric, ignoring that label correlations prevail asymmetric in real-world applications. To overcome the problem, we propose a novel Relative Entropy and PageRank-based Classifier Chains for multi-label classification (REPCC). Firstly, to completely capture label correlation, REPCC exploits relative entropy to estimate the asymmetric correlation between pairs of labels, filter out those pairs of labels with lower correlation, and construct a label correlation matrix. Secondly, by treating each label as a node, a directed graph is generated based on the filtered label correlation matrix. Finally, drawing an analogy between labels and web pages, the idea of PageRank, commonly used to measure the importance of web pages, is employed to rank labels in the directed graph and establish a classifier chain. Experimental results on ten public datasets from various domains show that REPCC outperforms state-of-the-art methods in CC.

Over the past few decades, the study of leadership theory has expanded across various disciplines, delving into the intricacies of human behavior and defining the roles of individuals within organizations. Its primary objective is to identify leaders who play significant roles in the communication flow. In addition, behavioral theory posits that leaders can be distinguished based on their daily conduct, while social network analysis provides valuable insights into behavioral patterns. Our study investigates five and six types of social networks frequently observed in different organizations. This study is conducted using datasets we collected from an IT company and public datasets collected from a manufacturing company for the thorough evaluation of prediction performance. We leverage PageRank and effective word embedding techniques to obtain novel features. State-of-the-art performance is obtained using various statistical machine learning methods, graph convolutional networks (GCN), automated machine learning (AutoML), and explainable artificial intelligence (XAI). More specifically, our approach can achieve state-of-the-art performance with an accuracy close to 90% for leaders identification with data from projects of different types. This investigation contributes to the establishment of sustainable leadership practices by aiding organizations in retaining their leadership talent.

Social media is a digital environment where users openly share their opinions and engage in debates and discussions on various topics. Social media has amassed an enormous quantity of accessible data due to its constantly expanding and highly active user base. This data is a valuable resource for researching a variety of topics. An example of such a research problem is identifying user similarities by analyzing their data. This article explores combining comments, textual posts, and likes for comments to create keywords or tags and group the users. These tags are extracted from user data and utilized in constructing a tag network. The tag network facilitates the formation of communities consisting of users with similar interests. User grouping is achieved based on the tags extracted from the posts. The proposed methods employ TF-IDF (term frequency-inverse document frequency) and TextRank algorithms to extract the tags. Kernel diffusion determines the similarities between tags in the tag network. Finally, an aggregation-hierarchical clustering algorithm is employed to group social media users based on these tags. [ABSTRACT FROM AUTHOR]

Purpose – This study analyzed Korea's relations table through network analysis. In particular, among the centralities, eigenvector centrality, PageRank centrality and degree were used. The author studied which network characteristics affected the value-added rate. Design/methodology/approach – A network analysis method was used. Findings – It is the inward relationship that affects the value-added ratio of Korea's industries and the outward relationship has less influence. In particular, the inward relationship not only acts as a cost but also has an effect on the rate of added value recently. Research limitations/implications – Since the three years of 2010, 2015 and 2019 are the target, the data are somewhat insufficient to generalize. Practical implications – As for the value-added ratio of an industry, input is more important than output (sales). Therefore, where the input is received is very important. Social implications – It is possible to increase the understanding of the determinants of the value-added rate of Korean industries. Originality/value – (1) It was clarified which side is inward or outward in determining the industry in Korea. (2) The relationship between PageRank, eigenvector centrality and degree was analyzed in Korean cases. (3) Input is a cost and acts to increase added value.

Many ranking algorithms and metrics have been proposed to identify high-impact papers. Both the direct citation counts and the network-based PageRank-like algorithms are commonly used. Ideally, the more complete the data on the citation network, the more informative the ranking. However, obtaining more data on citation relations is often costly and challenging. In some cases, obtaining the citation counts can be relatively simple. In this paper, we look into using the additional citation counts but without additional citation relations to form more informative metrics for identifying high-impact papers. As an example, we propose enhancing the original PageRank algorithm by combining the local citation network with the additional citation counts from a more complete data source. We apply this enhanced method to American Physical Society (APS) papers to verify its effectiveness. The results indicate that the proposed ranking algorithm is robust against missing data and can improve the identification of high-quality papers. This shows that it is possible to enhance the effectiveness of a network-based metric calculated on a relatively small citation network by including only the additional data of the citation counts, without the additional citation relations.

One of the modern directions of obtaining information for making informed decisions is the analysis of data from open Internet sources, the analysis of media containing hundreds of thousands of publications. It is critically important not only to obtain reliable information, but also the time needed to obtain and analyze it. The purpose of the research in this work is the development and testing of a complex methodology for quickly building a prototype of a system for complex analysis of thematic sites. A technology of interconnected methods, methodologies, and tools for building a graph database, a knowledge graph, data analysis using methods and mo dels of machine learning with the provision of analytical results to users has been created. The main task of this work is to use these technologies to analyze data from well-known world sites in order to build a prototype of a systems for complex analysis of data from Internet sources.

Currently, high-performance computing environments are facing challenges such as limited resources and an increasing number of users. In order to improve the utilization of environmental resources, this paper proposes a high-performance hybrid computing architecture based on big data processing technology, which is constructed on the basis of an HDFS distributed system combined with MapReduce framework and GPU virtualization technology. The PageRank algorithm is utilized to evaluate the performance of rack nodes in the high-performance computing environment, and the evaluation results are applied to design an improvement strategy for task allocation and scheduling through the MapReduce framework. A division function is introduced to dynamically divide the Reduce data, and an approximate sampling method based on sampling information is proposed to guide the setting of the number of Reduce. The IB algorithm is used to cluster the labeled files, and a rack-aware strategy is designed based on HDFS to achieve resource load balancing. The MapReduce-based task allocation scheduling scheme has a reduction in job execution time of up to 39.83% compared to delayed scheduling. The dynamic partitioning design can achieve data load balancing by partitioning 5.382% of the groups and migrating 1.207% of the KVs if the data skew is 1.0. Dynamic balancing of environmental resources and resource scheduling optimization in high-performance computing environments can be achieved through the use of big data processing techniques.

In the digital era, the proliferation of new media, facilitated by digital platforms, has positioned new media marketing as a predominant marketing strategy. This study introduces a novel node ranking algorithm, leveraging the principles of PageRank, and proposes the UIEM-CMR influence assessment model derived from it. Utilizing data extracted from microblogging platforms and Facebook, this model evaluates the influence of various new media marketing nodes. Subsequently, the ten most influential nodes within the microblogging dataset were analyzed to examine the correlation between node influence in new media marketing and both the thematic distribution of new media marketing and the community distribution of followers. The findings reveal that the MR value of node influence, as calculated by the model, is impacted by the combined effects of AR and BR values, with these values alone identified as pivotal. The influence ranking, when solely based on AR and BR values, tends to diverge significantly from real-world scenarios. Nodes ranked highly under popular new media marketing topics exhibit a substantially greater probability of distribution compared to lower-ranked or randomly chosen nodes. Conversely, the distribution probabilities among influential nodes under less popular (cold) new media marketing topics show no significant disparity. Nodes with high influence ratings tend to attract followers from at least two distinct communities. The UIEM-CMR model, as developed in this study, proves to be an effective tool for assessing the influence of new media marketing strategies.

Creating policy measures is the final step in the process of e-learning roadmap development. Policy measures can be seen as long-term activities that need to be implemented and constantly upgraded to achieve strategic goals. For resource allocation, it is useful to prioritize policy measures. Prioritization can be implemented using multi-criteria decision-making methods. This paper analyzes policy measures in the Maldives National University's e-learning roadmap using the social network analysis process (SNAP), which includes the analytic hierarchy process (AHP), the decision-making trial and evaluation laboratory (DEMATEL), and the PageRank centrality. In policy measure evaluation, there were more than 20 participants: persons with managerial functions at the Maldives National University (MNU) (deans, heads of departments) and persons in lecturer and researcher positions. By using the AHP, participants prioritized policy measures with respect to their importance to them. By using the DEMATEL, participants identified and prioritized policy measures with respect to their effect on other measures. Finally, by using the SNAP, it was possible to determine the prioritization list for resource allocation since it aggregates the aspects of the policy measures, their importance, and their effect on other measures. [ABSTRACT FROM AUTHOR]

There are errors in multi-source uncertain time series data. Truth discovery methods for time series data are effective in finding more accurate values, but some have limitations in their usability. To tackle this challenge, we propose a new and convenient truth discovery method to handle time series data. A more accurate sample is closer to the truth and, consequently, to other accurate samples. Because the mutual-confirm relationship between sensors is very similar to the mutual-quote relationship between web pages, we evaluate sensor reliability based on PageRank and then estimate the truth by sensor reliability. Therefore, this method does not rely on smoothness assumptions or prior knowledge of the data. Finally, we validate the effectiveness and efficiency of the proposed method on real-world and synthetic data sets, respectively. [ABSTRACT FROM AUTHOR]

Image sentiment analysis technology can predict, measure and understand the emotional experience of human beings through images. Aiming at the problem of extracting emotional characteristics in art appreciation, this article puts forward an innovative method. Firstly, the PageRank algorithm is enhanced using tweet content similarity and time factors; secondly, the SE-ResNet network design is used to integrate Efficient Channel Attention (ECA) with the residual network structure, and ResNeXt50 is optimized to enhance the extraction of image sentiment features. Finally, the weight coefficients of overall emotions are dynamically adjusted to select a specific emotion incorporation strategy, resulting in effective bimodal fusion. The proposed model demonstrates exceptional performance in predicting sentiment labels, with maximum classification accuracy reaching 88.20%. The accuracy improvement of 21.34% compared to the traditional deep convolutional neural networks (DCNN) model attests to the effectiveness of this study. This research enriches images and texts' emotion feature extraction capabilities and improves the accuracy of emotion fusion classification. [ABSTRACT FROM AUTHOR]

Ranking algorithms play a crucial role in information technologies and numerical analysis due to their efficiency in high dimensions and wide range of possible applications, including internet ranking, scientometrics, and systemic risk in finance (SinkRank and DebtRank). The traditional approach to internet ranking goes back to the seminal work of Sergey Brin and Larry Page, who developed the initial method PageRank (PR) in order to rank websites for search engine results based on linear algebra rules. But how robust is this method in times of rapid internet growth? Recent works have studied robust reformulations of the PageRank model for the case when links in the network structure may vary; that is, some links may appear or disappear, influencing the transportation matrix defined by the network structure. In this article, the authors make a further step forward, allowing the network to vary not only in links but also in the number of nodes. The authors focus on growing network structures and develop methods for ranking of networks uncertain both in size and in structure. Internet ranking algorithms play a crucial role in information technologies and numerical analysis due to their efficiency in high dimensions and wide range of possible applications, including scientometrics and systemic risk in finance (SinkRank, DebtRank, etc.). The traditional approach to internet ranking goes back to the seminal work of Sergey Brin and Larry Page, who developed the initial method PageRank (PR) in order to rank websites in search engine results. Recent works have studied robust reformulations of the PageRank model for the case when links in the network structure may vary; that is, some links may appear or disappear influencing the transportation matrix defined by the network structure. We make a further step forward, allowing the network to vary not only in links, but also in the number of nodes. We focus on growing network structures and propose a new robust formulation of the PageRank problem for uncertain networks with fixed growth rate. Defining the robust PageRank in terms of a nonconvex optimization problem, we bound our formulation from above by a convex but nonsmooth optimization problem. Driven by the approximation quality, we analyze the resulting optimality gap theoretically and demonstrate cases for its reduction. In the numerical part of the article, we propose some techniques which allow us to obtain the solution efficiently for middle-size networks avoiding all nonsmooth points. Furthermore, we propose a coordinate-wise descent method with near-optimal step size and address high-dimensional cases using multinomial transition probabilities. We analyze the impact of the network growth on ranking and numerically assess the approximation quality using real-world data sets on movie repositories and on journals on computational complexity. Funding: This work was supported by the Austrian Science Fund [Grant J3674-N26]. Supplemental Material: The online appendices are available at https://doi.org/10.1287/opre.2022.2298. [ABSTRACT FROM AUTHOR]

In the field of deep matching, a large amount of noisy data in Chinese long texts affects the matching effect. Most long-form text matching models use all text data indiscriminately, which results in a large amount of noisy data, and thus the PageRank algorithm is combined with Transformer to filter noise. For sentence-level noise detection, after calculating the overlap rate of words to evaluate the similarity, a sentence-level relationship graph is constructed and filtered by using the PageRank algorithm; for word-level noise detection, based on the attention score in Transformer, a word graph is established, then the PageRank algorithm is executed on graph, combined with self-attention weights, to select keywords to highlight topic relevance, the noisy words are filtered sequentially at different layers in the module, layer by layer. In addition, during the model training, PolyLoss is applied to replace the traditional binary Cross-Entropy loss function, thus reducing the difficulty of hyperparameter tuning. Finally, a better filtering strategy is proposed and experiments are conducted to verify it on two Chinese long-form text matching datasets. The result shows that the matching model based on the noise filtering strategy of this paper can better filter the noise and capture the matching signal more accurately. [ABSTRACT FROM AUTHOR]

In this study, we explore how transformer models, which are known for their attention mechanisms, can improve pathogen prediction in pastured poultry farming. By combining farm management practices with microbiome data, our model outperforms traditional prediction methods in terms of the F1 score—an evaluation metric for model performance—thus fulfilling an essential need in predictive microbiology. Additionally, the emphasis is on making our model’s predictions explainable. We introduce a novel approach for identifying feature importance using the model’s attention matrix and the PageRank algorithm, offering insights that enhance our comprehension of established techniques such as DeepLIFT. Our results showcase the efficacy of transformer models in pathogen prediction for food safety and mark a noteworthy contribution to the progress of explainable AI within the biomedical sciences. This study sheds light on the impact of effective farm management practices and highlights the importance of technological advancements in ensuring food safety.

In this paper, a new multi-parameter iterative algorithm is proposed to address the PageRank problem based on the multi-splitting iteration method. The proposed method solves two linear subsystems at each iteration by splitting the coefficient matrix, considering therefore inner and outer iteration to find the approximate solutions of these linear subsystems. It can be shown that the iterative sequence generated by the multi-parameter iterative algorithm finally converges to the PageRank vector when the parameters satisfy certain conditions. Numerical experiments show that the proposed algorithm has better convergence and numerical stability than the existing algorithms. [ABSTRACT FROM AUTHOR]

Consider chemical elements as a system, we create an undirected chemical network with 99 elements and 1916 edges from Chemspider, a website that provide search engines to collect compounds. Using this network and the network that we used in our previous work with 97 elements and 2198 edges, we found that RootedPageRank, a link prediction tool in complex network, can be used to predict potential binary compounds, because the changing trend of PageRank probability of each element in these networks all follow the periodic law, despite of the difference of scale of these networks. The accuracy test indicates that at least 7 among top 10 predicted compoundss in one network can be verified using the compoundss in the other network or in other chemical database, proving that this method can be used to provide guidance in finding potential binary compounds, suggesting that we can study chemical properties from the view of complex network. [ABSTRACT FROM AUTHOR]

Computational drug repositioning approaches are important, as they cost less compared to the traditional drug development processes. This study proposes a novel network-based drug repositioning approach, which computes similarities between disease-causing genes and drug-affected genes in a network topology to suggest candidate drugs with highest similarity scores. This new method aims to identify better treatment options by integrating systems biology approaches. It uses a protein-protein interaction network that is the main topology to compute a similarity score between candidate drugs and disease-causing genes. The disease-causing genes were mapped on this network structure. Transcriptome profiles of drug candidates were taken from the LINCS project and mapped individually on the network structure. The similarity of these two networks was calculated by different network neighborhood metrics, including Adamic-Adar, PageRank and neighborhood scoring. The proposed approach identifies the best candidates by choosing the drugs with significant similarity scores. The method was experimented on melanoma, colorectal, and prostate cancers. Several candidate drugs were predicted by applying AUC values of 0.6 or higher. Some of the predictions were approved by clinical phase trials or other in-vivo studies found in literature. The proposed drug repositioning approach would suggest better treatment options with integration of functional information between genes and transcriptome level effects of drug perturbations and diseases. [ABSTRACT FROM AUTHOR]

Abstract With the development of avionics industry, it is difficult for traditional combat equipment node evaluation method to meet our requirements under complex combat system. This paper presents a method of node importance evaluation which is suitable for modern avionics field and can be used for reference in other combat fields. In order to make better use of the different features of the node itself and the different connections between nodes, we use TOPSIS algorithm to model the characteristics of the node itself, and PageRank to measure the interdependence of all nodes. Therefore, a novel node contribution evaluation algorithm based on TOPSIS and PageRank is proposed in this paper. In addition, after the evaluation of node contribution, we found that there was also a functional relationship between the operational information entropy in the whole graph and the contribution of these nodes. On this basis, information entropy evaluation algorithm of the overall combat map is further proposed. After a lot of experiments, the reliability of our algorithm is evaluated on the indexes of the node's destruction-resistant performance and information transfer efficiency. Compared with the traditional universal algorithm, our proposed algorithm shows more interpretable and robust results in the field of avionics.

It is difficult for users to understand the complex cloud product information for product selection. Using this information to recommend satisfactory cloud products is a challenge. Previous studies focused on similar information of users and products while neglecting relevance; therefore, they could not create recommendation approaches that account for functional dependencies among cloud products. To overcome this challenge, this study proposes a cloud product set recommendation model based on a hierarchical knowledge graph (KG) with a pre-post correlation of product functionality. There are two main contributions: First, we constructed a cloud product functionality and performance KG using the dependency information of layers and entities to represent complicated pre-post logical connections. The KG was designed according to the cloud service model. Second, we designed an improved PageRank algorithm to obtain the importance weight for each functionality and performance, which replaces the original average method with the proportion of connection weight. We considered the release time of the functionality, launch time of the product, and last update time of the product as crucial factors in the recommendation score to reflect the importance of the functionality and current development stage of the product. Finally, our method recommended a product set based on the weighted scores from the above results. In addition, we constructed a cloud product functionality dataset containing 339 functionalities. The experimental results show that the proposed method can generate a closely related set of products, leading to improved accuracy and higher satisfaction compared to mainstream methods.

Our objective is to survey recent results concerning the evolution of random networks and related extreme value statistics, which are a subject of interest due to numerous applications. Our survey concerns the statistical methodology but not the structure of random networks. We focus on the problems arising in evolving networks mainly due to the heavy-tailed nature of node indices. Tail and extremal indices of the node influence characteristics like in-degrees, out-degrees, PageRanks, and Max-linear models arising in the evolving random networks are discussed. Related topics like preferential and clustering attachments, community detection, stationarity and dependence of graphs, information spreading, finding the most influential leading nodes and communities, and related methods are surveyed. This survey tries to propose possible solutions to unsolved problems, like testing the stationarity and dependence of random graphs using known results obtained for random sequences. We provide a discussion of unsolved or insufficiently developed problems like the distribution of triangle and circle counts in evolving networks, or the clustering attachment and the local dependence of the modularity, the impact of node or edge deletion at each step of evolution on extreme value statistics, among many others. Considering existing techniques of community detection, we pay attention to such related topics as coloring graphs and anomaly detection by machine learning algorithms based on extreme value theory. In order to understand how one can compute tail and extremal indices on random graphs, we provide a structured and comprehensive review of their estimators obtained for random sequences. Methods to calculate the PageRank and PageRank vector are shortly presented. This survey aims to provide a better understanding of the directions in which the study of random networks has been done and how extreme value analysis developed for random sequences can be applied to random networks. [ABSTRACT FROM AUTHOR]

Graph coloring is widely used to parallelize scientific applications by identifying subsets of independent tasks that can be executed simultaneously. Graph coloring assigns colors the vertices of a graph, such that no adjacent vertices have the same color. The number of colors used corresponds to the number of parallel steps in a real-world end-application. Therefore, the total runtime of the graph coloring kernel adds to the overall parallel overhead of the real-world end-application, whereas the number of the vertices of each color class determines the number of the independent concurrent tasks of each parallel step, thus affecting the amount of parallelism and hardware resource utilization in the execution of the real-world end-application. In this work, we propose a high-performance graph coloring algorithm, named ColorTM, that leverages Hardware Transactional Memory (HTM) to detect coloring inconsistencies between adjacent vertices. ColorTM detects and resolves coloring inconsistencies between adjacent vertices with an eager approach to minimize data access costs, and implements a speculative synchronization scheme to minimize synchronization costs and increase parallelism. We extend our proposed algorithmic design to propose a balanced graph coloring algorithm, named BalColorTM, with which all color classes include almost the same number of vertices to achieve high parallelism and resource utilization in the execution of the real-world end-applications. We evaluate ColorTM and BalColorTM using a wide variety of large real-world graphs with diverse characteristics. ColorTM and BalColorTM improve performance by 12.98 × and 1.78 × on average using 56 parallel threads compared to prior state-of-the-art approaches. Moreover, we study the impact of our proposed graph coloring algorithmic designs on a popular end-application, i.e., Community Detection, and demonstrate the ColorTM and BalColorTM can provide high-performance improvements in real-world end-applications across various input data given. [ABSTRACT FROM AUTHOR]

Computational drug repositioning approaches are important, as they cost less compared to the traditional drug development processes. This study proposes a novel network-based drug repositioning approach, which computes similarities between disease-causing genes and drug-affected genes in a network topology to suggest candidate drugs with highest similarity scores. This new method aims to identify better treatment options by integrating systems biology approaches. It uses a protein-protein interaction network that is the main topology to compute a similarity score between candidate drugs and disease-causing genes. The disease-causing genes were mapped on this network structure. Transcriptome profiles of drug candidates were taken from the LINCS project and mapped individually on the network structure. The similarity of these two networks was calculated by different network neighborhood metrics, including Adamic-Adar, PageRank and neighborhood scoring. The proposed approach identifies the best candidates by choosing the drugs with significant similarity scores. The method was experimented on melanoma, colorectal, and prostate cancers. Several candidate drugs were predicted by applying AUC values of 0.6 or higher. Some of the predictions were approved by clinical phase trials or other in-vivo studies found in literature. The proposed drug repositioning approach would suggest better treatment options with integration of functional information between genes and transcriptome level effects of drug perturbations and diseases.

With the development of avionics industry, it is difficult for traditional combat equipment node evaluation method to meet our requirements under complex combat system. This paper presents a method of node importance evaluation which is suitable for modern avionics field and can be used for reference in other combat fields. In order to make better use of the different features of the node itself and the different connections between nodes, we use TOPSIS algorithm to model the characteristics of the node itself, and PageRank to measure the interdependence of all nodes. Therefore, a novel node contribution evaluation algorithm based on TOPSIS and PageRank is proposed in this paper. In addition, after the evaluation of node contribution, we found that there was also a functional relationship between the operational information entropy in the whole graph and the contribution of these nodes. On this basis, information entropy evaluation algorithm of the overall combat map is further proposed. After a lot of experiments, the reliability of our algorithm is evaluated on the indexes of the node's destruction-resistant performance and information transfer efficiency. Compared with the traditional universal algorithm, our proposed algorithm shows more interpretable and robust results in the field of avionics. [ABSTRACT FROM AUTHOR]

We consider a network-based framework for studying causal relationships in financial markets and demonstrate this approach by applying it to the entire U.S. stock market. Directed networks (referred to as “causal market graphs”) are constructed based on publicly available stock prices time series data during 2001–2020, using Granger causality as a measure of pairwise causal relationships between all stocks. We consider the dynamics of structural properties of the constructed network snapshots, group stocks into network-based clusters, as well as identify the most “influential” market sectors via the PageRank algorithm. Interestingly, we observed drastic changes in the considered network characteristics in the years that corresponded to significant global-scale events, most notably, the financial crisis of 2008 and the COVID-19 pandemic of 2020.

Creating policy measures is the final step in the process of e-learning roadmap development. Policy measures can be seen as long-term activities that need to be implemented and constantly upgraded to achieve strategic goals. For resource allocation, it is useful to prioritize policy measures. Prioritization can be implemented using multi-criteria decision-making methods. This paper analyzes policy measures in the Maldives National University’s e-learning roadmap using the social network analysis process (SNAP), which includes the analytic hierarchy process (AHP), the decision-making trial and evaluation laboratory (DEMATEL), and the PageRank centrality. In policy measure evaluation, there were more than 20 participants: persons with managerial functions at the Maldives National University (MNU) (deans, heads of departments) and persons in lecturer and researcher positions. By using the AHP, participants prioritized policy measures with respect to their importance to them. By using the DEMATEL, participants identified and prioritized policy measures with respect to their effect on other measures. Finally, by using the SNAP, it was possible to determine the prioritization list for resource allocation since it aggregates the aspects of the policy measures, their importance, and their effect on other measures.

Computation ranking algorithms are widely used in several informatics fields. One of them is the PageRank algorithm, recognized as the most popular search engine globally. Many researchers have improvised the ranking algorithm in order to get better results. Recent research using Rayleigh Quotient to speed up PageRank can guarantee the convergence of the dominant eigenvalues as a key value for stopping computation. Bolzano's method has a convergence character on a linear function by dividing an interval into two intervals for better convergence. This research aims to implant the Bolzano algorithm into Rayleigh for faster computation. This research produces an algorithm that has been tested and validated by mathematicians, which shows an optimization speed of a maximum 7.08% compared to the sole Rayleigh approach. Analysis of computation results using statistics software shows that the degree of the curve of the new algorithm, which is Rayleigh with Bolzano booster (RB), is positive and more significant than the original method. In other words, the linear function will always be faster in the subsequent computation than the previous method.

In this paper, a generalized full orthogonalization method (GFOM) based on weighted inner products is discussed for computing PageRank. In order to improve convergence performance, the GFOM algorithm is accelerated by two cheap methods respectively, one is the power method and the other is the extrapolation method based on Ritz values. Such that two new algorithms called GFOM-Power and GFOM-Extrapolation are proposed for computing PageRank. Their implementations and convergence analyses are studied in detail. Numerical experiments are used to show the efficiency of our proposed algorithms.

Abstract The MetaCore commercial database describes interactions of proteins and other chemical molecules and clusters in the form of directed network between these elements, viewed as nodes. The number of nodes goes beyond 40 thousands with almost 300 thousands links between them. The links have essentially bi-functional nature describing either activation or inhibition actions between proteins. We present here the analysis of statistical properties of this complex network applying the methods of the Google matrix, PageRank and CheiRank algorithms broadly used in the frame of the World Wide Web, Wikipedia, the world trade and other directed networks. We specifically describe the Ising PageRank approach which allows to treat the bi-functional type of protein–protein interactions. We also show that the developed reduced Google matrix algorithm allows to obtain an effective network of interactions inside a specific group of selected proteins. In addition to already known direct protein–protein interactions, this method allows to infer non trivial and unknown interactions between proteins arising from the summation over all the indirect pathways passing via the global bi-functional network. The developed analysis allows to establish an average action of each protein being more oriented to activation or inhibition. We argue that the described Google matrix analysis represents an efficient tool for investigation of influence of specific groups of proteins related to specific diseases.

The identification of patient-specific cancer driver genes plays a crucial role in the development of personalized cancer treatment and drug development. Several computational methods have been proposed for identifying patient-specific cancer driver genes, most of which rank driver genes ac-cording to scores calculated from various gene or protein network information. In this paper, we propose a machine learning model for more accurate identification of patient-specific cancer driver genes. The training data for the proposed model is composed of the gene vectors, which indicate the impacts that one gene can have on or receive from all the genes. The gene vector is patient-specific, in other words, one gene can have many gene vectors from many cancer patients. To make gene vectors, first a patient-specific gene network is built using the gene expression data of each cancer patient and gene regulatory network, then modified PageRank is applied to the patient-specific gene network to make the impact matrix, from which gene vectors can be extracted. We used the Random Forest model to train gene vectors to find and discriminate patterns that show how known driver genes affect, or are affected by, other genes. The proposed model was tested through cross validations and independent tests using different sets of known cancer driver genes and six cancer types from The Cancer Genome Atlas (TCGA) data, and showed higher F-scores than existing patient-specific driver gene identification algorithms. The majority of predicted driver genes were rare, and F-scores calculated with these rare genes are higher than or comparable to those of frequently identified driver genes.

As an important part of distributed graph computing, graph partitioning has been widely studied. However, the majority of the existing approaches to distributed graph partitioning barely take into consideration the relationship between the partition strategy and the graph algorithm's characteristics. The current graph partitioning strategies are on the basis of the graph structure, and the performance is diverse as running different graph algorithms. In this paper, considering the characteristics of a graph algorithm, we propose a distributed graph partitioning framework that can decide which partitioning strategy is better based on the program analysis. We also design a triangle-based partitioning strategy which benefits several graph algorithms such as the triangle counting. In addition, to reduce the number of shuffling operations in PageRank algorithm, we redesign the PageRank algorithm which considers the corresponding hash partitioning. The experimental results show that our adaptive graph partitioning framework implemented on Graphx surpasses the original one, especially for the triangle counting and PageRank, no matter on real-world graphs or synthetic graphs. [ABSTRACT FROM AUTHOR]

The types of pharmaceutical products include cosmetics and drugs. Some of the pharmaceutical products comprise a mix of drugs and herbs without considering their interaction effects. Drug-herb interactions (DHIs) refer to the interactions between conventional drugs and herb medicines. However, the available information on DHIs is scattered because it has heterogeneous databases and website resources, apart from some of the paid or subscribed databases. Easy access to information on DHIs would allow researchers to explore more. Therefore, this study proposes improvements in the focus web crawler to collect DHIs information from the heterogeneous resources on the Internet, present priority levels of a resource link through anchor text and URLs, and traversing the link with the aid of depth. The improved focused crawler was tested on two algorithms namely the Breadth-First Search (BFS) and PageRank. Information of DHIs crawled 4,744 herbals from the focus web crawler. The accuracy values for Chinese Med Digital Projects and MedlinePlus were 98% for PageRank and 71% for BFS. Additionally, a focused web crawler may gather more relevant web pages in the same amount of time as a wide crawler. Hence, the proposed crawler may successfully gather DHIs on the web in response to the user queries. [ABSTRACT FROM AUTHOR]

Neural network-based encoder–decoder (ED) models are widely used for abstractive text summarization. While the encoder first reads the source document and embeds salient information, the decoder starts from such encoding to generate the summary word-by-word. However, the drawback of the ED model is that it treats words and sentences equally, without discerning the most relevant ones from the others. Many researchers have investigated this problem and provided different solutions. In this paper, we define a sentence-level attention mechanism based on the well-known PageRank algorithm to find the relevant sentences, then propagate the resulting scores into a second word-level attention layer. We tested the proposed model on the well-known CNN/Dailymail dataset, and found that it was able to generate summaries with a much higher abstractive power than state-of-the-art models, in spite of an unavoidable (but slight) decrease in terms of the Rouge scores. [ABSTRACT FROM AUTHOR]

By considering a weighted inner product, an adaptive generalized Arnoldi (GArnoldi) method was constructed by [13] for computing PageRank. In order to accelerate the adaptive GArnoldi method, this paper proposes a new method by using the power method with extrapolation process based on Google matrix’s trace (PET) as an accelerated technique of the adaptive GArnoldi method. The new method is called as GArnoldi-PET method, whose implementation and convergence analysis are discussed in detail. Numerical experiments are used to illustrate the effectiveness of our proposed method.

This research presents a cost-effective approach toward an enhancement of Blockchain consensus targeting enhanced data security and privacy in the Industry 4.0 scenario by incorporating an enhanced reputation score at a cost-effective approach. Cellular Automata (CAs)-based PageRank validation (i.e., an extended PageRank) algorithm is incorporated towards consensus design in Blockchain Trilemma to facilitate cost-effectiveness in Industry 4.0 scenario. A detailed investigation is carried out to explore the true potential for CAs-based PageRank validation towards computation of Reputation score in Blockchain Trilemma. Further investigations and experimental analysis for the proposed approach towards computation of Reputation scores explore the efficiency of the proposed CAs-based PageRank validation-based approach over several existing approaches e.g., PageRank, NCDawareRank, and HodgeRank-based approaches. Additionally, energy efficiency in the proposed approach is ensured as low-energy consumption may be achieved with CAs-based physical modeling.