Abstract

Network inference, which is the reconstruction of the connectivity structure of the network from recorded data, is essential for broadening our knowledge of physical, biological, and chemical systems. While data-driven network inference algorithms have made significant strides in recent years, determining how much data is required so that the inferred network topology faithfully mirrors the underlying network is an essential but often overlooked subject. In this manuscript, we present a statistical method to determine if the recorded data carries sufficient variability to ensure an accurate reconstruction of the true network topology. Our approach leverages parametric confidence intervals to determine the bounds of true connection strengths, which subsequently enable the uncertainty quantification of inferred connectivity. The proposed technique is validated by examining the noisy data generated from networks of Kuramoto and Stuart-Landau oscillators. The method is also applied to experimentally obtained data generated from an electrochemical oscillator network, where we find that the data sufficiency technique can successfully predict the accuracy of the inferred network.

Original languageEnglish
JournalIEEE Transactions on Network Science and Engineering
DOIs
StateAccepted/In press - 2025

Keywords

  • Confidence intervals
  • Network inference
  • Network topology
  • Nonlinear oscillators

Fingerprint

Dive into the research topics of 'Uncertainty Quantification of Network Inference with Data Sufficiency'. Together they form a unique fingerprint.

Cite this