中国邮电高校学报(英文) ›› 2019, Vol. 26 ›› Issue (1): 40-48.doi: 10.19682/j.cnki.1005-8885.2019.0005

• Networks • 上一篇    下一篇

Research on performance of complex networks based on principal component analysis

徐春霞,齐小刚,刘立芳   

  1. 西安电子科技大学
  • 收稿日期:2018-07-03 修回日期:2018-10-23 出版日期:2019-02-26 发布日期:2019-02-27
  • 通讯作者: 徐春霞 E-mail:934663448@qq.com
  • 基金资助:
    基于可信计算和大数据的无线传感器网络安全技术研究及验证;宁波市自然科学基金项目;复杂电子系统仿真重点实验室基础研究基金

Research on Performance of Complex Networks Based on Principal Component Analysis

  • Received:2018-07-03 Revised:2018-10-23 Online:2019-02-26 Published:2019-02-27

摘要: For complex networks, their effectiveness and invulnerability are extremely important. With the development of complex networks, how to evaluate the effectiveness and invulnerability of these networks becomes an important research topic. The relationship among many influencing factors is very complicated, so it is essential to confirm the weighting coefficient of these influencing factors. Principal component analysis (PCA) is proposed to evaluate the performance of complex networks. It can improve one-sidedness of the single evaluation index and select different evaluation models according to different complex networks, which make the evaluation result more accurate. Performance of complex networks can be predicted according to comprehensive evaluation model. To verify the rationality and validity of this method, several small-world networks with different probability values and scale-free network are chosen to evaluate the network performance. Finally, simulation results show that PCA can be applied to performance evaluation of complex networks.

关键词: complex networks, network topology, principal component analysis, performance prediction

Abstract: For complex networks, their effectiveness and invulnerability are extremely important. With the development of complex networks, how to evaluate the effectiveness and invulnerability of these networks becomes an important research topic. The relationship among many influencing factors is very complicated, so it is essential to confirm the weighting coefficient of these influencing factors. Principal component analysis (PCA) is proposed to evaluate the performance of complex networks. It can improve one-sidedness of the single evaluation index and select different evaluation models according to different complex networks, which make the evaluation result more accurate. Performance of complex networks can be predicted according to comprehensive evaluation model. To verify the rationality and validity of this method, several small-world networks with different probability values and scale-free network are chosen to evaluate the network performance. Finally, simulation results show that PCA can be applied to performance evaluation of complex networks.

Key words: complex networks, network topology, principal component analysis, performance prediction