Network topology inference with estimated node importance

¹ Adaptive Networks and Control Lab, Department of Electronic Engineering, and Research Center of Smart Networks and Systems, School of Information Science and Engineering, Fudan University - Shanghai 200433, China
² MOE Frontiers Center for Brain Science, Institutes of Brain Science, Fudan University - Shanghai 200433, China

^(a) lix@fudan.edu.cn (corresponding author)

Received: 31 August 2020
Accepted: 1 March 2021

Abstract

In real life, the actual topology of a network is often difficult to observe or even unobservable, which seriously limits our analysis and understanding of such networks. How to accurately infer the network structure from easily observed data is extremely urgent. In this letter, we try to improve the inference accuracy by introducing the heterogeneity of nodes during the network reconstruction, and propose a novel method to estimate the importance of nodes directly from the spreading results. The results on both synthetic and empirical data sets show that our algorithms can effectively improve the inference accuracy, especially when the observed data is insufficient.