Reconstructing networks from dynamics with correlated noise
Publication in refereed journal

香港中文大學研究人員
替代計量分析
.

其它資訊
摘要Reconstructing the structure of complex networks from measurements of the nodes is a challenge in many branches of science. External influences are always present and act as a noise to the networks of interest. In this paper, we present a method for reconstructing networks from measured dynamics of the nodes subjected to correlated noise that cannot be approximated by a white noise. This method can reconstruct the links of both bidirectional and directed networks, the correlation time and strength of the noise, and also the relative coupling strength of the links when the coupling functions have certain properties. Our method is built upon theoretical relations between network structure and measurable quantities from the dynamics that we have derived for systems that have fixed point dynamics in the noise-free limit. Using these theoretical results, we can further explain the shortcomings of two common practices of inferring links for bidirectional networks using the Pearson correlation coefficient and the partial correlation coefficient.
著者Tam HC, Ching ESC, Lai PY
期刊名稱Physica A: Statistical Mechanics and its Applications
出版年份2018
月份7
日期15
卷號502
出版社ELSEVIER SCIENCE BV
頁次106 - 122
國際標準期刊號0378-4371
電子國際標準期刊號1873-2119
語言英式英語
關鍵詞Network reconstruction,Correlated noise, structure and dynamics
Web of Science 學科類別Physics, Multidisciplinary;Physics

上次更新時間 2020-29-10 於 02:04