README.md

In racdale/sindyr: Sparse Identification of Nonlinear Dynamics

sindyr: Tools for Sparse Identification of Nonlinear Dynamics with R

sindyr is an R library for inferring ordinary differential equations from raw data, based on SINDy by Brunton et al. (2016). The library also extends this method in several ways, including a sliding window tool that generates descriptive measures of a system from SINDy.

If you use the code, toss us a shout out by citing:

Dale, R. & Bhat, H. S. (2018). Equations of mind: Data science for inferring nonlinear dynamics of socio-cognitive systems. Cognitive Systems Research, 52, 275-290.

Under "dale-bhat-materials," we show several example uses of sindyr, based on the manuscript cited above. For a quick start, consider perusing the reconstruction of the Lorenz attractor's equations using sindyr here.

Installation

sindyr is now available on CRAN:

install.packages('sindyr')

On occasion this version on GitHub will go out of alignment before we do a CRAN update. If this happens you can install from the repository this way:

install.packages('devtools')

devtools::install_github('racdale/sindyr')

Notes

sindyr works with R >= 3.4.

sindyr depends on the following packages: arrangements, matrixStats, igraph, and pracma.

References:

Brunton, S. L., Proctor, J. L., & Kutz, J. N. (2016). Discovering governing equations from data by sparse identification of nonlinear dynamical systems. Proceedings of the National Academy of Sciences, 113(15), 3932-3937.

racdale/sindyr documentation built on May 6, 2024, 6:26 p.m.