China
Africa
Explore chapters and articles related to this topic
Big Graph Analytics: Techniques, Tools, Challenges, and Applications
Published in Mohiuddin Ahmed, Al-Sakib Khan Pathan, Data Analytics, 2018
Dhananjay Kumar Singh, Pijush Kanti Dutta Pramanik, Prasenjit Choudhury
A major problem with Katz centrality arises when a node with high centrality is connected to many nodes and then all these nodes get high centrality. In many cases, however, this is not so significant if a node is only one among the many that are linked. To solve this problem, one can divide the passed centrality value by the number of outgoing links (out-degree) from that node such that each connected neighbor from the source node gets a fraction of the source node’s centrality. The PageRank centrality [34] of a node vi is defined as Cp(vi)=α∑j=1nAj,iCp(vj)djout+β
Structure of networks
Published in Karthik Raman, An Introduction to Computational Systems Biology, 2021
clearly justifying the name of eigenvector centrality. Here, λ is the largest eigenvalue of the adjacency matrix. Other popular centrality measures derived from eigenvector centrality include Katz centrality [9] and Google PageRank [10]. Named after Google co-founder Lawrence Page, PageRank ranks websites in Google's search results, based on the number and quality of links to a given page.
Network Science
Published in Paul L. Goethals, Natalie M. Scala, Daniel T. Bennett, Mathematics in Cyber Research, 2022
All measures mentioned above are defined in the context of undirected networks. When dealing with directed networks, these definitions need to be reformulated accordingly. Other centrality measures in the literature include Katz centrality, PageRank, hubs and authorities, etc. Due to the technicality of the arguments involved therein, we omit the details and refer the reader to Newman (2010).
Finding influential users in microblogs: state-of-the-art methods and open research challenges
Published in Behaviour & Information Technology, 2022
Umar Ishfaq, Hikmat Ullah Khan, Shahid Iqbal, Mohammed Alghobiri
Network functionality in OSNs is an important aspect for influence ranking. However, functional properties are network-dependent since each social network offers different interaction possibilities. Drakopoulos et al. (2017) proposed an analytical framework which extends the performance of an influence metric using probabilistic tools from multiple fields such as information theory, psychometrics and data mining. The proposed framework consists of influence metrics which can be classified into two different categories i.e. first order and higher order. First-order metrics utilises only the account-specific information of users. On the other hand, higher-order metrics derive the social influence of a user's account as a function of the influence of follower accounts. Resultantly, the digital influence of a user is effectively captured. The computation of higher-order metrics is based on Katz centrality (Katz 1953) and TunkRank (Daniel Tunkelang 2015). The former exploits the network structure whereas the latter employs the functional aspects.
Malicious accounts detection from online social networks: a systematic review of literature
Published in International Journal of General Systems, 2021
Imen Ben Sassi, Sadok Ben Yahia
Graph centrality and properties based approaches have computed several social network features to be used on the detection model, principally the degree of density, the degree centrality, the betweenness centrality, the in-degree, the out-degree, the reciprocity, the bi-directional links ratio, the eigenvector centrality, the Katz Centrality, and the Load centrality. The best results for Sybil detection have been obtained by Mulamba, Ray, and Ray (2018), where the authors have selected the set of features that can be able to discriminate benign nodes from malicious nodes, i.e. the core number, the weighted degree-degree centrality, the weighted core-degree centrality, the degree-coherence centrality, the core-coherence centrality, the edge volume centrality, the average degree centrality, and the average clustering centrality. Two datasets have been used to gauge the performance of the proposed method that achieved for the precision, the recall, the F-Measure, and the AUC with the Facebook dataset.
The roles of supply network centralities in firm performance and the moderating effects of reputation and export-orientation
Published in Production Planning & Control, 2020
Antonio K. W. Lau, Yuya Kajikawa, Naubahar Sharif
Second, this study is the first in the SCM literature to adopt a new measure of corporate reputation, PageRank. PageRank explores the reputation or influencing the power of focal firms in supply networks based on supplier-buyer contractual relationships (Zhang, Zhang, and Guo 2015). It has been used frequently for assessing authors’ reputations (Yan and Ding 2011; Ding et al. 2009), articles (Ma, Guan, and Zhao 2008) and journals (Cheang et al. 2014; Bollen, Rodriguez, and Van de Sompel 2006), as well as for examining influence on website visits (Glick et al. 2014) and purchasing decisions (Dhar et al. 2014). In SCM, PageRank can measure the degree to which a firm receives incoming supply chain links from other highly influential firms. It is a variant of common measures of reputation such as eigenvector centrality or Katz centrality. These common measures have a potential problem that firms linked to many reputable supply chain partners may be wrongly recognized as reputable, from which PageRank can prevent (Sharma et al. 2019; Zafarani, Abbasi, and Liu 2014) However, PageRank has not been used in supply chain or organizational reputation literature. This study thus extends the literature by introducing a new way to measure organizational reputation based on supply chain partners (Boivie, Graffin, and Gentry 2016) and answers the call to introduce new social network measures in the SCM field (Wichmann and Kaufmann 2016).