Singhal Bharat, Vu Minh, Zeng Shen, Li Jr-Shin
Department of Electrical and Systems Engineering, Washington University in St. Louis, St. Louis, MO, USA.
IFAC Pap OnLine. 2023;56(2):10089-10094. doi: 10.1016/j.ifacol.2023.10.879. Epub 2023 Nov 22.
Decoding the connectivity structure of a network of nonlinear oscillators from measurement data is a difficult yet essential task for understanding and controlling network functionality. Several data-driven network inference algorithms have been presented, but the commonly considered premise of ample measurement data is often difficult to satisfy in practice. In this paper, we propose a data-efficient network inference technique by combining correlation statistics with the model-fitting procedure. The proposed approach can identify the network structure reliably in the case of limited measurement data. We compare the proposed method with existing techniques on a network of Stuart-Landau oscillators, oscillators describing circadian gene expression, and noisy experimental data obtained from Rössler Electronic Oscillator network.
从测量数据中解码非线性振荡器网络的连接结构是理解和控制网络功能的一项困难但必不可少的任务。已经提出了几种数据驱动的网络推理算法，但在实践中，通常所考虑的充足测量数据这一前提往往难以满足。在本文中，我们通过将相关统计与模型拟合过程相结合，提出了一种数据高效的网络推理技术。所提出的方法在测量数据有限的情况下能够可靠地识别网络结构。我们将所提出的方法与现有技术在斯图尔特 - 兰道振荡器网络、描述昼夜节律基因表达的振荡器以及从罗斯勒电子振荡器网络获得的噪声实验数据上进行了比较。
