閱讀全文 | |
篇名 |
Entropy and Structural-Hole Based Node Ranking Methods
|
---|---|
並列篇名 | Entropy and Structural-Hole Based Node Ranking Methods |
作者 | Chinenye Ezeh、Tao Ren、Yan-Jie Xu、Shi-Xiang Sun、Zhe Li |
英文摘要 | Several research works had been carried out to discover suitable algorithms to quantify node centralities. Among the many existing centrality metrics, only few consider centrality at the sub-graph level or deal with structural hole capabilities of pivot nodes. Research has proven the importance of sub-graph information in distinguishing influential nodes. In this work, two centrality metrics are proposed to distinguish and rank nodes in complex networks. The first metric called Subgraph Degree Information centrality is based on entropy quantification of a node’s sub-graph degree distribution to determine its influence. The second metric called Subgraph Degree and Structural Hole centrality considers a node’s sub-graph degree distribution and its structural hole property. The two metrics are designed to efficiently support weighted and unweighted networks. Performance evaluations were done on five real world datasets and one artificial network. The proposed metrics were equally compared against some classic centrality metrics. The results show that the proposed metrics can accurately distinguish and rank nodes distinctly on complex networks. They can equally discover highly influential and spreader nodes capable of causing epidemic spread and maximum network damage. |
起訖頁 | 1009-1017 |
關鍵詞 | Sub-graph degree、Entropy、Node ranking、Structural hole、Influential node |
刊名 | 網際網路技術學刊 |
期數 | 202109 (22:5期) |
出版單位 | 台灣學術網路管理委員會 |
DOI |
|
QR Code | |
該期刊 上一篇
| A Deep Learning Based Equalization Scheme for Bandwidthcompressed Non-orthogonal Multicarrier Communication |
該期刊 下一篇
| Using Machine Learning Approaches to Improve Ultra-Wideband Positioning |