The reconstruction of equivalent underlying model based on direct causality for multivariate time series

View article
PeerJ Computer Science

Main article text

 

Introduction

  • In the pursuit of underlying model identification, the transfer entropy method emerges as a pivotal tool for unveiling the topology of causality. Moreover, the conditional transfer entropy or conditional mutual information approach is harnessed to delineate the direct causality structure precisely. This approach can be further tailored to deduce a collection of fundamental elements, facilitating the reconstruction of an equivalent underlying model grounded in multivariate time series data.

  • The polynomial fitting method determines the coefficients and intrinsic order of the causality structure based on the foundational elements derived from the direct causality topology. This method serves as a means to accomplish the task of underlying model identification, all from the vantage point of multivariate time series, without any reliance on prior knowledge. The culmination of this process leads to the discernment of the expression characterizing the equivalent underlying model.

Algorithms

Entropy

Mutual information

Conditional mutual information

Transfer entropy

Conditional transfer entropy

Significance test

Polynomial fitting

Simulation Study

Case 1: simulation case

Case 2: Henon map chaotic time series

Conclusions

Supplemental Information

Supplemental Figures

DOI: 10.7717/peerj-cs.1922/supp-1

Equivalent underlying model data

Case 1 is a discrete-time dynamical system. Case 2 is a Henon map chaotic time series. For each case, the training dataset consists of 700 samples, while the testing dataset comprises 300 samples.

DOI: 10.7717/peerj-cs.1922/supp-3

Additional Information and Declarations

Competing Interests

The authors declare there are no competing interests.

Author Contributions

Liyang Xu conceived and designed the experiments, performed the experiments, analyzed the data, authored or reviewed drafts of the article, and approved the final draft.

Dezheng Wang conceived and designed the experiments, performed the experiments, analyzed the data, performed the computation work, prepared figures and/or tables, and approved the final draft.

Data Availability

The following information was supplied regarding data availability:

The simulated cases are available in the GitHub and Zenodo:

- https://github.com/Xu-Liyang/case-reconstruction.git.

- Wang, D. (2024). The reconstruction of equivalent underlying model based on direct causality for multivariate time series. Zenodo. https://doi.org/10.5281/zenodo.10668929.

Funding

This research was funded by the Science and Technology Research Program of Chongqing Municipal Education Commission of China (Grant Nos. KJZD-K202201901, KJQN202201109, and KJQN202101904). This work was also supported by the Innovation Research Group of Universities in Chongqing (Grant No. CXQT21035), and the Scientific Research Foundation of Chongqing Institute of Engineering (Grant No. 2020xzky05). Additionally, this work was supported by Natural Science Foundation of Chongqing (Grant No. CSTB2022NSCQ-MSX1419, cstc2021jcyj-msxmX0525) and the Scientific and Technological Research Key Program of Chongqing Municipal Education Commission (Grant No. KJZD-M202201901). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

623 Visitors 604 Views 34 Downloads

MIT

Your institution may have Open Access funds available for qualifying authors. See if you qualify

Publish for free

Comment on Articles or Preprints and we'll waive your author fee
Learn more

Five new journals in Chemistry

Free to publish • Peer-reviewed • From PeerJ
Find out more