EWSmethods (development version)

EWSmethods 1.3.1

Bug fix for tensorflow installation on Windows - downgraded to tensorflow 2.13.0

EWSmethods 1.3.0

Remove tidyr and dplyr dependencies but now require >= R 4.4.0. Temporary fix for EWSNet as requires full rewrite for tensorflow 12.6.1. Addition of the Granger causality method - Information Imbalance.

EWSmethods 1.2.5

Patch to ewsnet_reset to ensure it gracefully fails when users’ internet connection fails.

EWSmethods 1.2.4

Hot fix to multiJI and uniJI following discussions with rEDM maintainers. These functions now average over time-varying Jacobians in a window rather than simply returning the final time-varying Jacobian.

EWSmethods 1.2.1

Due to API changes in the dependency rEDM, all Jacobian index functions were non-longer functional. Hot fix to convert functions to new rEDM API.

EWSmethods 1.2.0

Introduce multiAR as an alternative to multiJI for estimating Jacobians. Additional bug fixes and warning messages for resilience metrics. Removed unnecessary and problematic bypass_reticulate_autoinit().

EWSmethods 1.1.2

First CRAN release following CRAN maintainer comments. ewsnet_*() functions no longer download files without asking permission.

EWSmethods 1.1.1

Response to CRAN maintainer comments regarding spelling errors and slow example run times.

EWSmethods 1.1.0

EWSmethods 1.0.0

First release

• Added a NEWS.md file to track changes to the package.